C-Day Spring 2018 Project Submission (Responses)
 Share
The version of the browser you are using is no longer supported. Please upgrade to a supported browser.Dismiss

 
View only
 
 
ABCDEFGHIJKLMNOPQRSTUVWXYZAAABAC
1
Timestamp
*******Submission deadline: March 25, 2018 ------ Notification of acceptance: Monday, April 2 ******* The project must be affiliated with KSU and is sponsored by KSU faculty/staff member. Non-KSU projects are not eligible to participate at C-Day.
KSU Supervisor or capstone/Internship course instructor NAME (Dr/Prof FirstName LastName)
Supervisor or instructor EMAIL
Type of project
Course prefix and number (e.g. CS 4893) If applicable
I (we) would like to demo the poster/game/project during Flash Session
Would like to nominate the project/game/project to be considered for the best in the category
Equipment Requirements
Poster/Game Title Limit 35 characters for the program (will be included in the program). Your actual title can be longer.
Author1 (FirstName LastName)
Author2 (FirstName LastName)
Author3 (FirstName LastName)
Author4 (FirstName LastName)
Author5 (FirstName LastName)
Author6 (FirstName LastName)
Project Description (will be included in the program)
Major CS/IT/SWE/CGDD
Research Question or Motivation
Materials and MethodsPreliminary Results
Team/Author(s) Contact Email
Comments
2
3/9/2018 14:57:47
Yes, the project is affiliated with KSU
Rongkai Guorguo@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
YesNo
Table with power ( (ONLY IF YOU HAVE A DEMO)
Time RunnerTimon Wood
Time Runner is a game that I am creating merely to showcase my skills as an environmental artist. The "game" aspect of this project is merely an extra component that isn't entirely needed to fulfill the purpose of this project. Unlike a lot of students majoring in Computer Game Development and Design, I do not have any aspirations of creating my own studio or creating my own AAA video game. Instead, I simply just want to be an environmental artist for whatever studio is willing to hire me.
BSCGDD
My biggest motivation for me creating this project is my desire to get hired by literally any gaming studio in the world. I could care less about making a profit for this game. I did tell myself, however, that I do want to create one passion project before graduating and taking on a full-time job.
I will be using 3ds Max to create the static meshes needed to create my environments, and I will be using Unreal Engine 4 as my primary gaming engine. As specified before, I am an environmental artist (and a programmer), but I am not an animator nor a character artist. I - most definitely - will be using assets from the Epic Game Marketplace. Of course, it goes without saying that I will be citing them as needed.
I started working on this "passion project" back in December, but couldn't commit to doing much with this project due to some time constraints involving both my job and my other courses. I do not expect this game to come out as a grandeur spectacle, but I will pour my heart and soul into creating a visually appealing game, just so that I'll get a job fresh out of college.
timonwood@gmail.com
I know that I seem nonchalant about this project, but on the contrary, I am quite anxious and excited to work on this. It may not be the best project to be showcased come presentation day, but please know that I will pour my heart and soul into making this game. Again, the main purpose of this game is to showcase my skills as an environmental artist - nothing more.
3
3/9/2018 21:29:25
Yes, the project is affiliated with KSU
Dr. Xieyxie2@kennesaw.eduGraduate researchNoNoEasel
Interactive Mathematics
Linda Vu
Teaching mathematical proofs on interactive mediums to improve problem solving and critical thinking skills.
MSCS
The motivation is to allow students free access to interactive learning
Survey and GeoGebraTBA
lvu@students.kennesaw.edu
4
3/13/2018 23:49:42
Yes, the project is affiliated with KSU
Dr. Selena Hejhe4@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
KSU GoAlbert LimAnthony SchellChase GodwinDayton ChamberlinNick WilsonPatrick Hilerio
The driving objective behind this project is to improve the student experience through a user friendly mobile app. Part of the problem with student involvement is poor communication of relevant information, and this app will be focused at presenting this in a visually pleasing, easy to access manner.

We are going to create an app called KSU Go. It will be initially available to Android users and feature information about the campus, such as a campus directory, emergency services, D2L, Owl Life (clubs and organizations), Handshake (career planning), social feeds, campus map, and BOB (KSU bus) route tracking.
BSCS
The purpose of this project is to deliver a mobile app that is focused on improving the student experience at KSU.
Android Studio, MySQL, Express.js, and AWS Debian.
A fully functioning android app that contains features including but not limited to campus directory, emergency services, D2L, Owl Life (clubs and organizations), Handshake (career planning), social feeds, campus map, and BOB (KSU bus) route tracking.
alim5@students.kennesaw.edu
5
3/14/2018 11:43:16
Yes, the project is affiliated with KSU
Dr. Selena Heshe4@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850NoYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
AutodrivePayton TurnageAdam ZolanKelly DuongKen ChongDuy Quach
Automated drive through kiosk to reduce restaurant operating costs and make ordering more accessible to speakers of non-dominant languages.
BSCS
Reduce staffing costs and make ordering more accessible to speakers of non-dominant languages.
Software
First simulation is promising.
paytonturnage@gmail.com
6
3/14/2018 15:48:17
Yes, the project is affiliated with KSU
Dr. Selena Heshe4@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
Smart Baby MonitorTaylor BoundsJayce ChenJordan MateenWilliam SimmonsBo WangMichael Wong
A baby monitor connected to an android application that utilizes facial recognition to determine when the baby is waking up and alert the parents.
BSCS
We have a sponsor who had an interest in a baby monitor that could recognize when the child was waking up. He wanted to see if we could create it as a proof of concept.
Materials: Raspberry Pi 3, Pi Camera, Sensors, Smartphone, Android Studio, and Firebase. We utilized the raspberry pi to send a live stream from the camera. We split the video stream to the android application and the firebase. Once received by the android application it utilizes a pre-trained algorithm to recognize if real time when the child is waking up or upset and creates an alert to warn the user. The algorithm is trained utilizing the TensorFlow framework before being loaded into the app.
Currently we have the android application recognizing the different states of a child. Granted it isn't perfect yet, we just need more images to retrain it on a larger scale to up the accuracy. We also have the raspberry pi streaming the footage and taking in the sensor data. We just need to split the video stream into two outputs; one to the Firebase and the other to be broadcasted. We also need t integrate the receiving of the stream with the analysis of it. Currently the android app analyzes the stream from the camera on the smartphone. Add
wsimmon9@students.kennesaw.edu
We are looking forward to presenting and seeing everyone else's amazing work!
7
3/14/2018 19:13:27
Yes, the project is affiliated with KSU
Dr Selena He
she4@kennesaw.view.usg.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4491NoYes
Table no power (ONLY IF YOU HAVE A DEMO)
Smart Pill BottleDylan ClarkBen Davenport-Ray
Creating a smart pill bottle that will allow a user to input his/her medication schedule and dosage then remind the user to take his/her medication at the correct time. It will also notify the user via push notification if he/she misses a dose.
BSCS
Both team members have ADHD. Thus, taking their medication at the correct time poses a significant hurdle.
Raspberry Pi Zero W, load cell, LEDs, mobile phone, OpenResty web toolkit, Linux server, SQLite DB
The hardware activates an alarm when it is time for the user to take his/her medication and sends a push notification to the user's phone if the alarm is not turned off (the user missed a dose).
dclark45@students.kennesaw.edu
8
3/15/2018 16:44:47
No, the project is not affiliated with KSU
Dr Selena Heshe4@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS4850NoNoEasel
9
3/15/2018 17:36:05
Yes, the project is affiliated with KSU
Dr. Selena Heshe4@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
Virtual EducationSourav DebnathAdi VrskicAndrea PinardiJeffrey DennisMarc Wise
Our project aims to add the level of immersion in an online class, by developing a suite of virtual reality scenes simulating a real physical classroom environment, for the betterment and enrichment of a student’s education.
BSCS
Can online education incorporate the same immersion and social interactions of physical classroom through virtual reality?
The project uses Google's daydream virtual reality platform. To use the application users will need a daydream compatible phone, a daydream view headset and a daydream controller.
The application allows users to log in, and shows them the classes they are enrolled in, the user then can enter a classroom and interact with other classmates. The user also can create a study room and invite other classmates to join, the user can request a virtual meeting with an instructor, and finally the user can join a virtual lounge to meet other virtual students and faculties.
sdebnath@students.kennesaw.edu
10
3/16/2018 16:42:56
Yes, the project is affiliated with KSU
Professor Reza Meimandi Parizi
rparizi1@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
SWE 4724YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
Cybriant SOC PortalAli AbdulelahCameron FiteZachary AuzenneBrett WarnerFrancisco EscalanteArjumand Fatima
In a Security Operations Center there is a need to certify that a task has been performed for future auditing. Unfortunately, there are many different tools on the market and only moderate integration into ticketing or task management platforms. This means that it is difficult to quantify the time it takes for Security Analysts to perform tasks. The ability to certify (or at least recognize) that a task has been completed and the time it takes enables an organization help identify efficiencies as well as opportunities for further employee training.
BSSWE
How can security operation center processes be made more efficient in terms of task completion and other logistics?
Materials used include two servers, one for development and the other for the production version of the application. Both servers are provided by Cybriant. A secured VPN connection to those servers is also provided by Cybriant. The software development environments used are NetBeans and MySQL Workbench. The two software applications used for virtual meetings are GlobalMeet and Google Hangout. The team is following the V-model development methodology to design, develop, and test the application.
The Software Requirements Specification has been finalized and approved by the client, along with the Software Design Document. Mock ups have been developed to display the team's vision of the product to the client. Cybriant approved of the mock ups, which has led the team to move on the development and implementation phase of the project.
aabdulel@students.kennesaw.edu
11
3/16/2018 18:45:45
Yes, the project is affiliated with KSU
Prof Selena Heshe4@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
ServiceAlertHeath GriffinAyoola BishiAaron HollandBrandon KelleySharon Xu
This is a vehicle maintenance web application and a mobile application. The web application can be used as a maintenance log and a map for nearby service centers. The mobile application has these functionalities as well as a GPS functionality to track a user's mileage and recommend maintenance when certain milestones are reached.
BAACS
Provide a user-friendly application to assist in vehicle maintenance scheduling and logging.
Visual Studio, Android Studio, Google Cloud Platform Database, Java, C#/.NET, XML, HTML, JavaScript, SQL
Web and Mobile Applications with Basic Funtionalities
hgriff12@students.kennesaw.edu
12
3/18/2018 10:28:42
No, the project is not affiliated with KSU
Dr. Selena Heshe4@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS4850YesYes
Table no power (ONLY IF YOU HAVE A DEMO)
13
3/18/2018 11:48:25
Yes, the project is affiliated with KSU
Dr. Ming Yang
myang8@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
IT 4983NoNo
Table no power (ONLY IF YOU HAVE A DEMO)
MoodleCourtney WalkerJessica HerrBrian LuongAdam LeeDoutimi Azu
The Distance Learning Center at Kennesaw State University is looking for an alternative to D2L in order to deliver its training programs to newly hired faculty. A previous Capstone group had already setup the Moodle server for testing. So, our goal for this semester has been to research and implement the various Reporting and Badge tools available in Moodle.
BSIT
The Distance Learning Center want to be able to use various cross-course reporting features within Moodle.
Moodle, Windows Server 2012, VMware.
Cross-course reporting features are available within Moodle.
cwalk162@students.kennesaw.edu
14
3/18/2018 14:49:36
Yes, the project is affiliated with KSU
Dr. Selena He she4@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850NoYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
IoT Voice & Gesture Control System
Deja Tyla JacksonZoe Cesar Richmond Mensah Nathaniel KleinAmanda Norris Khadijah Mahley
A gesture and voice recognition system that connects to AWS IoT core to publish a command. A robotic car is subscribed to the cloud and it receives the command and performs an associated action.
BSCS
Our Research Question is: How can the smart classroom be improved through the use of IoT connected systems? Our motivation is: Smart classroom environment where teachers can move freely from their podium and be able to connect to their computers through the use of IoT connected systems.
Raspberry Pi Robot, Raspberry Pi with SenseHAT Attachment, Amazon Web Services, Google Voice API
The Google Voice API works for our intended commands in a client environment. The Raspberry Pi is able to publish commands to the cloud and the robot is able to subscribe to the robot to retrieve these commands to control its motors. The preprocessing script successfully records sensor data and formats it correctly for labeling in the initial machine learning model. Once the user is involved, the gesture data will be accurately formatted for the final machine learning model, so that the results will be as accurate as possible.
djack133@students.kennesaw.edu
We would love to have both a table with power, as well as board for the poster to stand up on. We also will need Wi-fi connectivity to KSU Guest, if Wi-fi will be a problem we will have to know ahead of time to modify the project to connect to a hotspot. Thank you!
15
3/18/2018 20:56:49
Yes, the project is affiliated with KSU
Dr. Mingon Kang
mkang9@kennesaw.edu
Undergraduate research
NoYesEasel
Animal Detection Using R-CNN
John JajehMasood Abdul Salam
By applying regions on CNN features, R-CNN provides computer vision solutions for multiple-object detection. In our research, we are utilizing AlexNet’s pre-trained model in the Caffe framework to detect approximately 400 different animal species and are acclimating this work from KSU’s GPU server to the Android environment. After an individual downloads the application and an animal is detected, he/she can click on the animal, which will prompt Google to search the animal label. Essentially, this app will allow users to photograph unfamiliar (or familiar) animals for identification and better personal understanding.
BSCS
Our motivation is to learn about machine learning and computer vision and apply these concepts to create an animal-detection application.
Materials: GPU Server w/ AlexNet, Caffe, Python, Jupyter Notebook

Methods:
1. Acquire data for testing.
2. Create memory-mapped database file.
3. Create mean image values.
4. Test to acquire results for labels and scores. Modify code as necessary and repeat testing until desired results are found..
5. Acclimate into Android environment and grant Google Search capability.
6. Test and modify accordingly.
Tested AlexNet’s pre-trained model (1000 object detection) with a dataset of randomly selected images produced via Selective Search with R-CNN.
jjajeh@students.kennesaw.edu
16
3/18/2018 23:02:53
Yes, the project is affiliated with KSU
Dr. Selena HeShe4@kennesaw.eduGraduate researchNoYesEaselIOT with Hand GestureNidhibahen PatelNidhibahenPatel
Gesture is movement of the body parts especially face, hand and arms that induces a meaningful expression and gesture recognition identifies the gist of the gesture. Gesture recognition is helpful in planning a smart and effective human–computer interface. The utilizations of gesture acknowledgment ranges from communication via gestures through therapeutic recovery to virtual reality. In this paper, we give a survey on gesture recognition with accentuation of hand gestures. Existing difficulties and future research potential outcomes are additionally highlighted
MSCS
Gesture is movement of the body parts especially face, hand and arms that induces a meaningful expression and gesture recognition identifies the gist of the gesture. Gesture recognition is helpful in planning a smart and effective human–computer interface. The utilizations of gesture acknowledgment ranges from communication via gestures through therapeutic recovery to virtual reality. In this paper, we give a survey on gesture recognition with accentuation of hand gestures. Existing difficulties and future research potential outcomes are additionally highlighted
Raspberry Pi , Google API
Hand Gesture Recognition
npate303@gmail.com
17
3/19/2018 0:39:23
Yes, the project is affiliated with KSU
Dr Selena Heshe4@kennesaw.eduGraduate researchNoYesEasel
Training a Neural Network to Walk
Oscar GarciaSelena He Ph.D.
Using a novel neural network architecture navigate an arena to reach a target. This is an implementation step towards achieving an architecture capable of solving insight class problems (the ability to synthesize solutions to a novel problem without training those solutions).
MSCS
The goal of my thesis is to build a neural network that can navigate a puzzle used to test for insight (the ability to synthesize solutions to a novel problem). The implementation is challenging so a more structured approach to the design process must be employed similar to the method employed by Taguchi where a feature (a hyper parameter of the design) is varied slowly while the others are held constant to determine the impact to output failures since analyzing real neural networks mathematically for a converging solution is exceedingly difficult.
Using Caffe neural network architecture library on multiple computers, design and test multiple architectures and hyper-parameters to search for a neural network architecture that can successfully control an avatar in an arena. Upon success the simplified arena will be expanded to the arena proposed for the thesis. Then the ability to choose actions based on a decision process will be added as a step towards insight.
Developed four classifications of training failures and one for success, analyzing for correlation between the failures (or success).
ogarcia5@students.kennesaw.edu
18
3/19/2018 12:33:23
Yes, the project is affiliated with KSU
Prof Dennis Franklin
dfrank15@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CGDD 4814NoYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
MalwarriorJacob Jennings
Malwarrior is a VR rail shooter for the Samsung Gear VR on Android. It uses various strategies to limit VR sickness on a mobile VR platform.
BSCGDD
How can we make a mobile VR game while limiting the chance of VR sickness?
User-Friendly types of movement, Reference Points, high frame-rates, low-poly, 3D sound.
Successful, After strategies implemented, there was a decrease in VR related nausea in players.
gamingkaleidoscope@gmail.com
19
3/19/2018 13:41:11
Yes, the project is affiliated with KSU
Dr Dennis Franklin
dfrank15@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CGDD 4814NoYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
FirestormBrandon TrippeAlan HornerLetisha Miller
Firestorm is a role playing game with turn-based combat, nods to games of similar genres, and an interesting narrative of progressive revelation. The player begins in a locked room by themselves, not remembering anything. The only thing they can do is get into their virtual reality machine, the Fireball. Their journey begins in an MMO they play with other people, all who have their own motives and goals, but they will join a resistance and fight against their captors, in both the virtual and real worlds.
BSCGDD
We wanted to make an RPG with nods to other games and a good story.
Unity, Game DesignBeta version of game
btrippe@students.kennesaw.edu
20
3/19/2018 14:35:37
Yes, the project is affiliated with KSU
Dr. Sarah Northsnorth@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
Visualization of Big Data with AR
Sean SaffanRonald BrooksVictor Orellana
Matt Alyward (HS Intern)
Project Title: Visualization of Streamed Big Data
Through Augmented Reality
The purpose of this project is to design, develop and measure the effectiveness of an augmented reality (AR) application for the visualization of streamed data of network traffic. This application will be developed for the Microsoft HoloLens, with several AR models of data visualization to select from. The effectiveness of this application will be measured in an experiment.
BSCS
Research Question:
1. Is there any significant difference in efficiency in using AR to visualize data streaming compared to 2D displays?
2. Does augmented reality visualization improve one's ability to monitor and analyze network traffic patterns?

The experiments used an array of laptops and other electronic devices such as Samson S2 9.7”; 32 GB Android Tablet. The main assessments were completed with the Microsoft HoloLens Mobile Devices managing (MDM) for HoloLens. Software included selected Universal Windows Platform (UWP) Apps to install and set security configurations; Visual and Android Studio 2017; Unity software, a cross-platform game engine developed by Unity Technologies, to develop 3-dimensional modeling interface within the HoloLens environment.
The experimental instruments included a pre-experiment survey and post-experiment survey. Specific questions that were administered to the participants for each experiment.
In the first phase of this study, as described above, a model augmented reality application for the visualization of streamed data of network traffic was designed and developed. Several visualization prototypes were designed and tested using device specific user interface commands.
The next phase was to measure the effectiveness of the AR system. An experiment was designed and implemented as follows: 20 participants tested the model AR visualizations on the HoloLens. They were given pre and post experiment surveys to answer. The results were then graphed and analyzed.

Many participants originally stated that they believed the HoloLens would be relatively easy to use. Comprehensive surveys and observations showed that it was actually more difficult for the participants to use the HoloLens.
Participants were largely positive about the benefits of AR technology to visualize streaming big data both before and after testing the model. Preliminary results for comprehension of the data displayed ranged from average to strong. Post-experiment, a strong majority of participants expressed that benefits exist for incorporating AR technology in data visualization.
ssaffan1@students.kennesaw.edu
Thank you!
21
3/19/2018 15:22:13
Yes, the project is affiliated with KSU
Selena He
she4@kennesaw.view.usg.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4491-02YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
IoT Smart Coffee PotBrent RobinsonJustin VoorheesChip Gardner
Implementation of a smart coffee pot utilizing a raspberry pi, temperature sensor and load cell to display relevant data and notify others when a new pot of coffee is ready.
BSCS
Create an internet of things application utilizing the concepts of Fog computing
Coffee pot, Raspberry Pi Zero W, Temperature Sensor, Load Cell. Method, connect sensors to Raspberry Pi. Affix sensors to coffee pot in meaningful areas and display data in an easy to use web app.
Users are messaged in a slack channel when a new pot of coffee is ready. Web app displays current coffee brewing status along with remaining cups of coffee. Settings are able to be changed through the web app.
brobin79@students.kenensaw.edu
22
3/19/2018 15:31:34
Yes, the project is affiliated with KSU
Dr. D. Michael Franklin
mfranklin@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CGDD 4814YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
Autonomous Robot Sim Using Unity
Christian ChoDerek MartinJesse MandersDr. D. Michael Franklinchristiancho
Simulation of an autonomous mobile robot in a closed office environment using Unity. The simulation is realized as an actual autonomous real world robot. The project will utilize real world imaging from the robot integrated inside of the Unity environment. The simulation will also include modeling of sensors and the implementation of an artificial intelligence.
BSCGDD
How tightly can we integrate a highly detailed 3D robotic model with its real world analog.
Hardware: Computer, Arduino, Raspberry Pi, Robotics Hardware, VR Goggles. Software: Unity with custom scripting, Custom Arduino and Raspberry Pi code.
Fully functioning simulation with highly accurate modeling. Functioning autonomous mobile robot and integrated sensor fusion.
ccho4@students.kennesaw.edu
23
3/19/2018 15:44:44
Yes, the project is affiliated with KSU
Dr. Michael Franklin
dfrank15@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CGDD 4814YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
Neon SkylineErik ShefferJacob HollowayJarrod AriolaHunter Jones
Neon Skyline is a fast paced, futuristic, and gritty racing game. In the year 2095, Adam Reeds is an ex-racer, who recently retired and escaped from the underground. He arrives home on New Years to find his daughter kidnapped, with nothing left behind but a note, and a suspicious set of keys. With the help of his mechanic, Sam Rodgers, Adam Reeds will have to re-enter the racing scene, to find his daughter, and take her back. Neon Skyline features local multiplayer, procedural generated race courses, and crazy ramps and tracks to keep your heart racing as fast as the hover cars flying by you.
BSCGDD
We wanted to make a racing game, that branched from the normal stereotypes of the genre. For example, instead of normal cars, they hover- allowing us to build crazy tracks that defy physics. We also wanted to help bring back the local multiplayer genre, as most games have strayed from it.
We used Unity as our engine to build the game, and Autodesk Maya to create the models. Paint.Net to make the sprites, and Ableton and Audactiy to make the audio.
We have a working prototype of the game that shows all of our featured mechanics.
jacobholloway045@gmail.com
24
3/19/2018 15:58:10
Yes, the project is affiliated with KSU
Dr. Dennis Michael Franklin
dfrank15@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CGDD 4814YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
AscensionZachary BakerAnthony OnukwuliNikki Ashaka
An archaeologist seeking ancient secrets plummets down an underground passage. As he attempts to find his way out, he stumbles upon a darkness that can be contained. In the inky depths of the cavern, he comes upon relics that contain the spirits of the Fallen Ones. Each of them, lamenting and contrite, were helpless victims of the ruthless power they attempted to oppose. With the power of the relics, the lost excavator must free these tortured spirits from their bondage and redeem the sins of ancient evils.
BSCGDD
To provide an enjoyable gaming experience while telling an in-depth and interesting story.
Unity Game Engine, Paint.NET, a PC capable of running the game, and various other media making tools.
A working prototype of the game.
aonukwul@students.kennesaw.edu
25
3/19/2018 16:08:44
Yes, the project is affiliated with KSU
Dr. Tsai-Tien Tsengttseng@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
4850NoYesEasel
FASTA Parallelization using CUDA
Carrick McClainJacob CrumbleyConnor ToaleAlbert PhouangaphayEdesiri Imarah
This project is an optimization of the FASTA protein search using the CUDA language and GPUs to parallelize specific areas of the FASTA program suite produced by William Pearson at the University of Virginia
BSCS
Help improve research done by Dr. Tseng and the biology department at KSU
CUDA C, FASTA suite
FASTA suite runs on GPU server with no parallelization
jcrumbl3@students.kennesaw.edu
26
3/19/2018 16:11:40
Yes, the project is affiliated with KSU
Dr. Yong Shiyshi5@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS4850NoNo
Table with power ( (ONLY IF YOU HAVE A DEMO)
Carbon Link Linden HaynesMatthew WoolridgeDevin MitchellDana CohenMatthew Weise
Customer loyalty application developed for Carbon using the Net Promoter scoring system. With our app, merchants can have a better understanding of customer behavior by receiving immediate feedback from their customers.
BSCS
We wanted to work with Verifone and become familiar with their software.
We used the Verifone SDK, Carbon, and Android studio.
The Verifone SDK for Carbon is still in development. We supplied feedback on where it needs work.
mweise@students.kennesaw.edu
27
3/19/2018 17:20:47
Yes, the project is affiliated with KSU
Dr. Shiyshi5@kennesaw.edu
Service (e.g. internship, capstone project, student club)
CS 4850YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
SchedulitJohn Mwaniki
Phoenix Alexander-Wright
Veronica BenavidesWill Zimmerman
Schedulit is a software as a service for small businesses to manage a shared schedule, handle client appointments, assigning client appointments to employees, and provide businesses a portal whereby clients can schedule appointments with the service provider for available services a business provides.
BSCS
The main concern the application addresses with respect to the targeted
users is the need for automation of mundane tasks; In this case, scheduling.
Front-end: HTML, CSS, and JavaScript. Back-end: Ruby on rails. Database: PostgresSQL
Users can create a business account, add services available to clients, add schedule of when they are available to provide the service. With respect to clients, they can navigate to a business profile, view the calendar and available dates, select a service to view the availability(day and time), select the day and time they want to book an appointment, and enter their information. The business can then view all the appointments clients have booked on their account.
johnmwaniki@outlook.com
28
3/19/2018 17:24:10
Yes, the project is affiliated with KSU
Dr. Hejhe4@kennesaw.edu
Undergraduate research
CS 4491NoNoEasel
RF Waveform Manipulation
Daniel Young
Manipulate a radio frequency based smart device by analyzing the waveform signals transmitted/received, manipulating those signals to control the device.
BSCSMotivated by IoT
Transciever and a laptop
Proved possible to control OOK type devices
dyoung28@students.kennesaw.edu
29
3/19/2018 18:38:00
Yes, the project is affiliated with KSU
Dr. Selena Heshe4@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS4850/02NoNo
Table with power ( (ONLY IF YOU HAVE A DEMO)
Bibliography Management Tool
William PatrickRandolph WilsonMichael LinAvery SnyderRoberto Recinos
A tool that will allow users to easily manage their cited sources when writing a paper, and generate a work cited page in MLA/APA format when they have finished writing their paper.
BSCS
Keeping track of sources and writing a works cited page at the end of a paper is annoying.
We've created a JavaFX program for the user to use, and will be using Elasticsearch in the backend to allow users to search for sources among the 25million records offered by the Library of Congress.
We have a mostly-working client that can Add/Edit/Remove sources from a list, and give the user a works cited page that they can paste into their paper.
wpatric4@students.kennesaw.edu
30
3/19/2018 22:48:12
No, the project is not affiliated with KSU
Prof Selena Heshe4@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
NoNo
Table with power ( (ONLY IF YOU HAVE A DEMO)
31
3/20/2018 10:34:37
Yes, the project is affiliated with KSU
Dr. Selena Hejhe4@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850NoNo
Table no power (ONLY IF YOU HAVE A DEMO)
RockdoveBen Davenport-RayLily ZaychikJason BradleyKyle ButterworthJackie Huggins
Rockdove is a distributed, decentralized, peer-to-peer chat system that is built using Go, IPFS, and other technologies for better reliability, reduced cost of service, and increased privacy and security through the elimination of servers.
BSCS
What can be improved with messaging systems? Cost of running service, reliability, privacy and security - all of which can be improved with the elimination of servers!
Go, IPFS, HTML, CSS
The chat system works as we expected, albeit somewhat slow in the beginning, but more peers should help make it faster. We are working on improving speed and efficiency.
lzaychik@students.kennesaw.edu
32
3/20/2018 11:57:48
Yes, the project is affiliated with KSU
Yong Shiyshi5@kennesaw.edu
Academic courses graduate (e.g. capstones, innovative special topics projects, other course projects)
cs 4850NoYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
Male vs Female control using Epoc+
Eric RawlsFrancesco BenjaminAbdullah Alamri Ashley Archibald
Controlling Parrot's AR Drone 2.0 with Emotiv Epoc+, a brain controller. We've integrated the two technologies together and tested to see the differences in control between male and female use.
BSCS
Is there advantages and disadvantages between male and female use of the Emotiv
Emotiv Epoc+; Parrot AR Drone 2.0
There is a significant value shown between male and female users.
erawlsy@gmail.com
33
3/20/2018 15:32:46
Yes, the project is affiliated with KSU
Professor Selena He
she4@kennesaw.view.usg.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850YesYes
Table no power (ONLY IF YOU HAVE A DEMO)
EDP (Enterprise Delivery Planning)
Alexander PlattelWilliam ParishDavid McFallClayton HunsingerHeather RegoChasse Cossairt
The purpose of EDP will be to offer a simple SAP interface for mobile devices which will allow users to update data in SAP in real time from their mobile device. We will be building the initial instance of this tool for an auto parts supply company called IMC and it will be a resource for truck drivers to register the items they deliver and the selected payment method and amount paid (although we will currently not be taking payments through the application).
BSCS
Our sponsor is a car parts manufacturing company called IMC. Part delivery truck drivers for IMC are currently given a shipment of products to be delivered to several different customers. Each customer has a specific order request with a set of materials allotted to them. Sometimes during the delivery process, the customer either no longer finds need for the part, or there are short payments, early payments, and over-payments in cash or check form.
The problem is that drivers currently receive a sheet of paper with a checklist of materials and payment methods which they have to fill out. After they give this paper back to the warehouse manager, the manager gives the sheet to employees who manage SAP data entry. Then, they enter this data into SAP. This process to check the deliveries that were successfully and unsuccessfully completed and update it in the SAP system costs IMC over sixty-man hours a week. However, our mobile application seeks to eliminate that cost.
Mobile applications, SAP data processing modules, and a gateway interface to transfer information between SAP and the applications.
A successful interface for our sponsor.
acpcreation@gmail.com
34
3/20/2018 16:59:02
Yes, the project is affiliated with KSU
Donghyun Kim
dkim76@kennesaw.edu
Graduate researchNoYesEasel
Tree Topology for VANET Application
Joong-Lyul LeeJaejin HwangHyesung ParkDonghyun Kim
Apply topology control to reduce network latency and improve energy efficiency of VANET(Vehicular Ad-hoc Network)
MSCS
Reduce network latency using new tree topology vonstruction scheme
Applying tree topology algorithm, Graph Theory, spanning tree, k-Lifetime tree, Optimazation, k-LTTC
Mobile Adhoc Network tree with topology algorithm cannot applied on the VANET. Developing topology algorithm
jhwang6@students.kennesaw.edu
This work has been submitted in the 6th IEEE Annual International Workshop on Mission-Oriented Wireless Sensor and Cyber-Physical System Networking (MiSeNet 2018), a Workshop of IEEE INFOCOM 2018, Honolulu, HI, USA, April 15-19, 2018. I will also attend the workshop to present my work.
35
3/20/2018 17:59:05
Yes, the project is affiliated with KSU
Prof. Yong shiyshi5@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
4850NoNo
Table with power ( (ONLY IF YOU HAVE A DEMO)
Rare Event DetectionRakeem DurandKarim RattaniIhsan HashemRoberta Beaulieu
Surveillance cameras have become a part of our life. We see them in almost every corner of the street and outside restaurants. However, hiring people to monitor all the surveillance video is expensive. The purpose of our project is to implement an algorithm that will detect abnormal activity such as robbery, accident, etc. and alert the corresponding authorities.
BSCS
We wanted to explore state-of-the-art approaches that detect anomalous or suspicious events in surveillance videos.
Customized Neural Networks
A model that detects anomalous events from supervised learning
rdurand1@students.kennesaw.edu
36
3/20/2018 20:22:44
Yes, the project is affiliated with KSU
Professor Selena He
she4@kennesaw.view.usg.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS4850NoYesEaselTravelBotRyan DrummEric BellRyan SparksTate BoiceSydnei ChurchJeremy Taylor
A chatbot that enhances a user's travel experience by suggesting and providing items such as hotels, flights, events, local news or travel accommodations.
BSCS
Simplify the travel experience using emerging technology.
Chatbot hosted on any Android device.
Chatbot
rdrumm@students.kennesaw.edu
37
3/20/2018 20:41:24
Yes, the project is affiliated with KSU
Dr. Selena Heshe4@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850NoYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
Parking Enforcer Application
Dominic AjagbaChris NelsonJacques Farris Hansuk Song Zack Baker
The Parking Enforcer Application (PEA) is a web based application designed to identify vehicles based on their licence plate, and send that information over to the parking authority.
BSCS
We wanted to develop an application that could be used by both the school and the students.
Raspberry Pi 3 & Raspberry pi camera
N/A
dajagba@students.kennesaw.edu
38
3/20/2018 22:23:39
Yes, the project is affiliated with KSU
Drs. Hisham Haddad and Hossain Shahriar
hhaddad@kennesaw.edu
Graduate researchCS 7999NoYesEasel
HIPAA Compliance Analysis of EHRs
Maryam FarhadiDr. Hisham HaddadDr. Hossain Shahriar
Electronic Health Record (EHR) applications are digital versions of paper-based patient's health information. EHR applications are increasingly being adopted in many countries. There have been reported breaches of Protected Health Information (PHI) data stored by EHR applications. In many reported breaches, improper use of EHRs has resulted in disclosure of patient’s PHI data. Inefficient application design threatens the integrity of EHRs, which leads to fraud and endangering patient's health. The goal of our research is to identify HIPAA technical requirements, evaluate open source EHR applications for security vulnerabilities using open-source scanner tools and map identified vulnerabilities to HIPAA technical requirements.
MSCS
Many vulnerabilities remain undetected in current EHR applications. Fixing security vulnerabilities becomes more expensive when they are discovered later. To identify HIPAA breaches in EHR applications, we applied open source static analysis tools that are intended to reveal web security bugs. We could identify some security vulnerabilities in EHR applications and map them to HIPAA technical requirements.
We used three open source static analysis tools named RIPS, VulnHunter, and Progpilot which identifies vulnerabilities in PHP applications.
This initial case study shows that EHR applications suffer from implementation level vulnerabilities which may impact HIPAA security requirements.
mfarhad1@students.kennesaw.edu
39
3/20/2018 23:55:23
Yes, the project is affiliated with KSU
Dr. Jing (Selena) Heshe4@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850NoNo
Table with power ( (ONLY IF YOU HAVE A DEMO)
PibbleNick MainCharles BryantSean HaleHaley BearMark Petell
“Pibble” (the common nickname for pitbulls) is a chrome extension with the ability to scrape data websites to assist our sponsor, Labrador, it will also implement machine learning for the application to work on other sites. Labrador has sponsored this team to create an alternative to the manual method of data collection and entry currently used by their employees. The plugin solution includes the collection of data from two main sites: Edward’s and Fortune’s.
BSCS
Our motivation is derived from the need of our sponsor.
We utilize the chrome extension store, javascript, and machine learning.
As of the middle of this semester we have succesfully met Labrador's minimum requirements.
chale17@students.kennesaw.edu
The team members may not all be CS majors.
40
3/21/2018 11:42:08
Yes, the project is affiliated with KSU
Dr Reza Parizi
rparizi1@kennesaw.edu
Academic courses graduate (e.g. capstones, innovative special topics projects, other course projects)
SWE 7903YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
PDF DExtractorNeal HughartCharlie SoSarada Sundaresan
PDF DExtractor is a web application that allows th end-user to upload a zip file to the website. The system then unzips the zip file and returns a random PDF form from a stack of PDF forms within the zip file. The end-user selects the text and format of the destination file that they would like the system to extract for all of the PDF forms in the zip file. The end-user is then given the option to save the selection coordinates for future use if they have created an account on the PDF DExtractor web site.
MSSWE
How do we create a web application that performs batch data extraction for all PDF forms in a zip file based on end-user data selection?
We used the rapid prototyping process, Amazon AWS products, Tabula.py, Git hub for version control, and Google Docs for documentation control.
Current we have a functional web application that can unzip a zip file and perform batch data extraction on multiple PDFs based on user data selection.
nhughart@students.kennesaw.edu
41
3/21/2018 12:42:20
Yes, the project is affiliated with KSU
Prof Dennis Franklin
dfrank15@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CGDD 4814YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
VR BloodstreamRobert ForbesKalib CroneJoshua Skelton
A Virtual Reality Experience that sends players into a virtual lab and bloodstream environment. The goal of the project is to teach students how to read and examine the bloodstream of different animals in order to determine when they are sick or irregularities are happening based on their bloodstream cell counts.
CGDD
Motivation was to provide Veterinary Students with a way to understand animal bloodstreams in a more up close and personal experience. This lead into the desire to improve their understanding of Animal Bloodstream Information by providing a visible and interactive environment.
HTC Vive, Custom Built Desktop Computer, Base Stands.
Using testing methods of observing users while they experience the environment we are able to determine what factors in the game are working or not based on users misunderstanding information or becoming confused on what to do.
NAjskelton@uga.edu
42
3/21/2018 14:36:45
Yes, the project is affiliated with KSU
Selena Heshe4@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS4491YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
IoT Occupancy TrackingAlex GoogeJoshua SanyikaMarc Wise
The project uses two to four PIR motion sensors connected to a Raspberry Pi to detect motion going inside or outside of a doorway, hallway, etc. It then keeps a count of the number of occupants inside of the given room or area. The project also includes a web application that shows data in a easy to use manner for end users.
BSCS
Marta Trains sometimes are extremely full during peak times. This project could be used for Marta as a "mapping" for times they are busy, or for end users to help find a less full train car.
4 PIR Sensors, Raspberry Pi, and a Laptop. The sensors and Pi might have a custom case.
We are able to track motion in and out of a doorway, but currently quick motion (faster than a normal walking speed) might be missed.
alexcgooge1@gmail.com
We would like a spot by a hallway or door if possible since that is the premise of our project.
43
3/21/2018 15:30:13
Yes, the project is affiliated with KSU
Dr. Rongkai Guorguo@kennesaw.eduGraduate researchYesYes
Table no power (ONLY IF YOU HAVE A DEMO)
User Traffic Patterns Models
Nick MurphyDevan PatelDrew SavasDerek MartinDr. Rongkai
In this phase of our research, we have developed a method for modeling the shape and size of "walkable" pathways of an environment based on traffic data collected from Bluetooth beacons in that area.
MSCS
Our motivation is to develop environmental models that can self-update based on changes in traffic throughout those areas. Possible applications of these models include providing navigation guidance to people with visual impairments as they walk through high traffic, constantly changing environments such as subway stations, airports, or malls.
[Methods]: We installed beacons in three unique locations within the Atrium building. Multiple participants then walked through those areas, and their positions were recorded over a period of time. This data was then combined, clustered together, filtered to remove outliers, and finally used to generate a navigation mesh.

[Materials]: Bluetooth Low Energy Beacon, Android mobile devices, Android Studio, Unity.
We were able to construct navigation meshes for three distinct sections of the Atrium with user position data gleaned from Bluetooth Low Energy beacons.
nmurphy1@students.kennesaw.edu
Thank you for your consideration
44
3/21/2018 15:46:50
Yes, the project is affiliated with KSU
Dr. Donghyun Kim
dkim76@kennesaw.edu
Graduate researchNoNoEasel
Smart Contract of Ethereum
Euiseong Ko
Ethereum, which is one of the crypto-currencies based on blockchain technology, has been studied for smart contract. Furthermore, I introduce the vulnerabilities in smart contract of ethereum.
MSCS
explore a cutting-edge technology called blockchain
applied cryptography and cryptographic protocol
I have studied how Bitcoin works and other crypto-currencies works.
eko1@students.kennesaw.edu
45
3/21/2018 16:41:28
Yes, the project is affiliated with KSU
Dr. Ying Xieyxie2@kennesaw.eduGraduate researchCS7999NoNoEasel
Sign Language Recognition
Kshitij Bantupalli
Recognizing American Sign Language dynamically using Deep Learning models like 3DCNN and Recurrent CNN combined with Computer Vision.
MSCS
This project aims at developing a bridge between sign language speakers and non sign language speakers by using technology to translate sign language to readable text.
Python, Tensorflow, Keras, Theano, Kaggle
Pre - processing the data and developing the model for the deep learning. Currently testing 3DCNN. I plan to modify it to the project and maybe add a Recurrent CNN to improve performance.
kbantupa@students.kennesaw.edu
I will be submitting another project for my Machine Learning class, and it would be great if I could have my easel close to that project.
46
3/21/2018 17:46:51
Yes, the project is affiliated with KSU
Dr. Yong Shiyshi5@kennesaw.eduGraduate researchNoNoEasel
K-Nearest Neighbor on Hadoop
Arialdis JapaDaniel Brown
For this research, we implemented the K-Nearest Neighbor algorithm in a MapReduce paradigm and tested it on a Hadoop cluster.
MSCS
Improvement possibilities for the K-Nearest Neighbor algorithm for data analysis.
A personal computer and a Hadoop cluster provided by the school.
A K-Nearest Neighbor implementation running on a local computer. It currently predicts MNIST data with 90% accuracy using 7 neighbors.
ajapa@students.kennesaw.edu
47
3/21/2018 17:47:14
Yes, the project is affiliated with KSU
Dr. Yong Shiyshi5@kennesaw.eduGraduate researchNoNoEasel
Clustering Algorithms for MapReduce
Daniel BrownArialdis Japa
An implementation of big data clustering algorithms for use with the MapReduce framework and Apache Hadoop.
MSCS
What performance benefits will cloud-based clustering algorithms provide?
A personal computer and Apache Hadoop cloud server
Currently have a working version of a common kMeans clustering algorithm for MapReduce, and am working on a different custom clustering algorithm (likely density-based).
dbrow380@students.kennesaw.edu
48
3/21/2018 17:51:14
Yes, the project is affiliated with KSU
Dr Yong Shiyshi5@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850 YesYesEasel
IT Service Delivery Optimization
Estephanos MekbibOnesmus Mungaru Andrew SwitzerKiana Barber Manoj Jaiwant
This project deals with the automation of the IT Customer Service Delivery for Wellstar Health Care System.
BSCS
How to automate the IT Customer Service Delivery in Wellstar Health Care system?
By analyzing and evaluating data and time tickets.
Designing a framework or solution to the IT Delivery Services
asweitzer@gmail.com
49
3/21/2018 18:26:17
Yes, the project is affiliated with KSU
Yong Shiyshi5@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850NoYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
KSU Online Management System
Khai TranSon TranJeremy SloanTong ChenWei Chen
The purpose of the project is to help improve and removate our current KSU Online Library Management System. We are designing and implementing a fully-functioning web application about online library system for an academic institute. Our goal is to create an organized and user-friendly online library system that fulfills the needs of the students and the staffs of the library.
Computer Science
To help improve and removate our current KSU Online Library Management System.
PHP/Laravel, Bootstrap, MySQL, HTML, CSS
A user-friendly Online library System
ktran24@students.kennesaw.edu
50
3/21/2018 19:24:14
Yes, the project is affiliated with KSU
Prof Yong Shiyshi5@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850NoYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
Virtual Reality Locomotion
Adam HayesVojtech MartinekAnderson McGeeAlex Kimbell
Our project is a research project focusing on Virtual Reality(VR) Locomotion. We want to answer several questions regarding VR Locomotion, such as "Which locomotion do users find easiest to use", and "With which locomotion do users maintain better spatial awareness". We built a simulation using the Unity Engine to give users simple tests that will help answer those questions. We designed both an obstacle course for them to navigate and a simple room to test their spatial awareness.
BSCS
To find out which Virtual Reality Locomotion users find the most comfortable to use, and which ones they are able to maintain spatial awareness while using these locomotion.
We used a Virtual Reality Headset called the HTC Vive, it comes with two handheld controllers and two sensors on mounts to track where the headset is in a room. The Vive headset is hooked up to an MSI gaming laptop that is running the Unity Engine. Our simulation was built entirely using the Unity Engine software with the Steam VR package installed. We also used Github to sync and share our work among the group to make sure everyone was working on an updated copy and we could catch integration errors early in the process.
Due to internal testing among the group, while we build the simulation, we believe that the easiest locomotion for people to use will be the Teleportation locomotion. This is by far the easiest for us to use and we expect it will be the same result when we begin our formal testing.
ahayes61@students.kennesaw.edu
51
3/22/2018 8:57:28
Yes, the project is affiliated with KSU
Dr. Mingon Kang
mkang9@kennesaw.edu
Academic courses graduate (e.g. capstones, innovative special topics projects, other course projects)
CS7267NoYesEasel
Medical Image Classification
Jessica Rudd
This project will compare traditional convolutional neural network (CNN) with the latest improvements to CNN technology known as CapsNet for improved medical image classification.
BSCS
Early detection of lung cancer is crucial for giving patients the best chance at survival. Using high resolution lung segmentation images, our aim is to improve upon existing CNN algorithms for image classification and reduce the rate of false positive detection.
This project utilizes data from 2017 Kaggle Data Science Bowl, high resolution lung segmentation images. Python will be used for data preprocessing and algorithm development.
CapsNet has been shown to have improved image classification results compared with traditional CNN for the MNIST classification problem. It is suspected that similar results will be found in medical image data.
jrudd1@students.kennesaw.edu
52
3/22/2018 11:18:21
Yes, the project is affiliated with KSU
Professor Yong Shiyshi5@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
Secure Mobile Software Development
Dyanne MarquezUsman JamilJustin ComerMatthew HullTredarious Booker
We will show case different android security methods used to help prevent security vulnerability. Our software development consisted of three methods that we created to help test vulnerability, Broadcast(), startActivity(),startServices().
CS
What are some methods that we can use to help prevent security vulnerability?
Android Studios; Java; Methods of Data Leakage
Method preparations
tbooker7@students.kennesaw.edu
53
3/22/2018 13:41:42
Yes, the project is affiliated with KSU
Dawn Tatum
dtatum7@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CSE 4983YesYesEaselInternship at GTRICodi Sordelet
I am interning as a front end developer. I am working on a web application on the front end that allows the Family advocacy program of the US military to track instances of abuse within Army families. I am using ,JavaScript, angular, grunt, karma, Sass, bootstrap, git version control, JavaScript, and HTML5
BSIT
Apply knowledge through industry experience
N/A
I have fixed many of the UI bugs, tested the web application on several different browsers and added an UI error detection feature
csordele@students.kennesaw.edu
54
3/22/2018 15:57:53
Yes, the project is affiliated with KSU
Dr. Selena Heshe4@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS4850YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
Alexa Adventure Creator
Bryan HallSuat IskenderAlex CobbJacob ConleyBen NunnelleyZach Black
Our application is a "create your own narration adventure" game for Amazon Alexa. We use an Android mobile application to create the game data, then save it to a database, then Alexa queries the data and narrates it to the user.
BSCS
Dungeons and Dragons for Amazon Alexa
Android mobile application, Alexa SDK, Maven, Java, Amazon Web Services, and Amazon Echo
Android app creates game data, saves that data to a Firebase database, and Alexa reads the data from a Firebase database and narrates it to the user.
bhall74@kennesaw.students.edu
55
3/22/2018 16:27:54
Yes, the project is affiliated with KSU
Dr Yong Shiyshi5@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850NoNoEaselFMSnoteMohamed KabadShainu VazhathilFaisal Abdulkadir
Android application that merges three of the most utilized apps by students. The application, “FMSnote”, will merge Reminders, Notes and To-Do applications. The application helps the users track all the important dates and events in their every day lives.
BSCS
People constantly lose track of very important dates, and also forget very important information that they intended not to. FMSnote will help out in taking notes, making reminders and making a To-do list all in one.
Android Studio, Java, XML, android phone.
A fully working app that do the intended functionalities.
mkabad@students.kennesaw.edu
56
3/22/2018 18:24:21
Yes, the project is affiliated with KSU
Yong Shiyshi5@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
Verifone Support WebApp
Srija KukkadapuAhmad AlissaDylan ClarkKevin BurkeJosh Sisson
A WebApp for Verifone with an easy-to-use interface that answers questions that exist within the customer base of the company
BSCS
What is the fastest way that Verifone could answer any questions that their customers might have?
HTML, CSS, Javascript, etc
A WebApp developed for Verifone to answer any questions that their customers might have.
skukkada@students.kennesaw.edu
57
3/23/2018 1:01:29
Yes, the project is affiliated with KSU
Dr. Mingon Kang
mkang9@kennesaw.edu
Academic courses graduate (e.g. capstones, innovative special topics projects, other course projects)
CS 7267NoNoEasel
Machine Learning income prediction
Yash BrahmbhattNidhi Kudiri
The objective of our project is to predict whether a person earns an income over 50 thousands a year or not.
MSCS
Extra Credit | to determine average pay scale depending on key attributes | To derive Fair wages
Laptop, Python (language)
Yet to find
nkudiri@students.kennesaw.edu
58
3/23/2018 10:57:06
Yes, the project is affiliated with KSU
Dr. Mingon Kang
mkang9@kennesaw.edu
Academic courses graduate (e.g. capstones, innovative special topics projects, other course projects)
CS7267YesYes
Table no power (ONLY IF YOU HAVE A DEMO)
House Prices- Regression Techniques
Aakanksha GuptaAditi Kasat
There are several factors that influence the house prices, the most important among them is location and other factors are economic growth of the city, house space, parking space, community safety, number of bedrooms, patios and many more. A price also depends on supply and demand, a seller who is willing to wait till he gets the price he thinks is right and the buyer who is willing to pay for the price. A rational approach facilitated by machine learning can be very useful in predicting the house price. A large data set with 79 different features (like living area, number of rooms, location, etc.) along with their prices are provided for residential homes in Ames, Iowa. The challenge is to learn a relationship between the important features and the price of the home and then use the above relationship to predict the prices of a new set of houses. The goal for the project is to predict the sales price for each house(Id).
MSCS
To implement the learning of Machine Learning course; Implement Regression Technique in Real World Problem
Methodology:
1. Acquire the data set
a. A Dataset “Ames, Iowa: Alternative to the Boston Housing Data” will be used for the project.
2. Explore the data set
a. Description given in 3rd Bulletin
b. Main challenge is to predict the sale price, which will be stored in SalePrice column (i.e Target Variable).
3. Engineer and transform the features and the SalePrice (Target Variable)
a. Examine correlation between data features and target.
b. As we see that some data has missing or null values we have to deal with that too, as it is important to gain domain knowledge about missing data.
c. We also have some non-numeric data we have to encode that too, for classification
d. Try to make data Boolean
4. Build a mode
a. Build a linear model
b. Split the data into training set and hold training set, so that we can see how it can perform on other sets as well, as we have to avoid overfitting
5. Calculate predictions
The goal for the project is to predict the sales price for each house(Id).
akasat@students.kennesaw.edu
Experiment Design
1. Plotting Graphs will be an essential role, as we want to visualize the distribution of data, see pattern, see classification and regression is done correctly.
2. Perform Regression to check skewness and improve linearity of data.
3. Calculate the Prediction and Check the Accuracy.
4. Make Data look like Normal Distribution.
5. Transform Categorical non-numeric and null data to numeric data and fill data using Interpolation.
6. Engineer features via Feature Engineering, so that data is suitable for Deep Learning
7. Apply Linear Regression on model and fit the model for hold training set
8. Evaluate performance using Root Mean Squared Error using 10-fold Cross Validation
9. Make a csv with prediction of SalePrice for each observation and check the differences between test and train data.
59
3/23/2018 11:46:25
Yes, the project is affiliated with KSU
Meng Hanmhan9@kennesaw.edu
Undergraduate research
IT4490NoYesEasel
UAVs and commercial flights
Jeffrey Collar
This project analyzes the effect civilian and military unmanned aerial vehicles (UAVs) have on commercial airline flights.
BSIT
I work directly with military air traffic control (ATC) and with the increase in UAVs in the recent years, I wanted to see what kind of effect that was having on commercial flights.
Data was acquired from the FAA UAV registration, Military drone usage records, and commercial airline delays to see if there is a correlation. A control group was also to isolate the effect of only the UAVs.
I have not gathered sufficient results yet to make a determination.
jcollar@students.kennesaw.edu
60
3/23/2018 12:38:52
Yes, the project is affiliated with KSU
Dr. Yong Shiyshi5@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850YesYes
Table no power (ONLY IF YOU HAVE A DEMO)
"What's Next?!" Travel App
Delia SandersShivani PatelKevin Benitez SantosAniruddh KathiriyaDhruv Patel
Create a travel App that allows users to create a custom list with their desired locations of the chosen activities, based on the selected timeline.
The app also provides a custom map for the user to reach the locations of their activities in the least amount of time
BSCS
The motivation for building this app was born from the frustration one experiences when traveling to a new location and having available time to use on activities, but due to unfamiliarity with the new place, planning for a pastime can be challenging.
The “What’s Next?! Travel app represents the senior project capstone project, during the Spring semester 2018, of group number 17, composed of Aniruddh Kathiriya,Dhruv Patel,Delia Sanders,Kevin Benitez Santos and Shivani Patel.
In creating the application, the methods used involved Android Studio for front end, Java for the backend functionality implementation, MySQL database information storage and android mobile devices for the application access and testing.
We created a User Interface with sign-up, login, questions to be answered and timeline selection, list creation with a map of the routes for the establishments on the list and sign out.
The interface connects to the database and changes can be made, in real time access of the application, through an Android mobile phone.
dsande41@students.kennesaw.edu
61
3/23/2018 15:23:41
Yes, the project is affiliated with KSU
Yong Shiyshi5@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850NoNoEaselWellstar IT OptimizationAndrew HubeliSarah HansberryDante TaylorAustin BarnettMichael Little
The Wellstar IT department receives upwards of 20000 tickets per month. The demand for IT services must be met while managing costs and resources. Many of the tickets could be resolved through automation or self-service.
BSCS
What processes in IT are most common and can be automated?
Data Analysis, Interviews, ServiceNow
Preliminary top 5 processes from both interviews and data analysis
ahubeli@students.kennesaw.edu
62
3/23/2018 15:34:39
Yes, the project is affiliated with KSU
Professor Dawn Tatum
dtatum7@kennesaw.edu
Service (e.g. internship, capstone project, student club)
CSE 4983NoNoEasel
Southern Company Internship
Andrew Evans
I started as an intern at Southern Company's Energy Innovation Center in February 2017. I was hired on as a software engineering intern and was expected to complete tasks within this field, as well as assist other employees in day-to-day operations. During my time at Southern Company, I have been able to participate on many amazing projects that dealt with Blockchain Technology, IoT platforms, AR/VR applications, 3D printing and Smart Cities of Atlanta.
BSSWE
To gain a deeper insight into the technology and innovation of a public utility
N/A
I have found that the Energy Innovation Center is here to foster a culture of innovation throughout Southern Company and to invent ideas and programs that provide a better service to their customers and community.
aevan119@students.kennesaw.edu
63
3/23/2018 17:34:56
Yes, the project is affiliated with KSU
Prof. Diana Rabahdrabah@kennesaw.edu
Undergraduate research
NoYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
Self-Schooling TabletPauline Sawadogo
The “Self-Schooling Tablet” is designed to be used as a learning support device for the millions of out of school children. The self-Schooling Tablet is connected to two trays.
Starting from the bottom, the first tray contains sand. The second tray is of two folds, the first is a homework sheet with markers & pencils container and the second is a white board.
Beginners, starts with watching video lessons on the tablet, follow instructions, and practice writing on the sand, the white board and then completing homework sheet for mentors to review.
Advanced learners will watch videos or PowerPoint lessons, practice on white board if applicable and then complete homework for mentors to review.
BASIT
According to UNESCO, 270 million children in the world are not schooled. My country, Burkina Faso, is among the top ten illiterate country in the world. As a former elementary school teacher, now graduating in IT from KSU, I am thrilled to use my education in technology to address the children illiteracy issue in the world & mainly in my country.
I will be using the Self-Schooling tablet, an educative support device that I designed, to demonstrate how it could be used to address millions of children illiteracy in the world.
-Handmade prototype of “Self-Schooling Tablet”.
-Draft of the “Supervised Self-Schooling program” procedure.
pysawadogo@gmail.com
According to Frederick Douglas:" it is easier to build strong children than to repair broken men". C day is an opportunity for me to start this noble journey of providing education for out of school children around the world.
64
3/23/2018 17:44:58
Yes, the project is affiliated with KSU
Dr Mingon Kang
mkang9@kennesaw.edu
Academic courses graduate (e.g. capstones, innovative special topics projects, other course projects)
CS 7267NoYesEasel
Predicting Crime Patterns
Kshitij BantupalliShade AlabsaBill Hawkes
The project aims at analyzing and understanding the pattern between crime in Atlanta and using that data to predict future occurrences of crime to help protect people and/or to also use Police resources judiciously.
MSCS
Given the data of crime in a major city, is it possible to find patterns using machine learning techniques to apply a preventative approach rather than a reactionary approach against crime?
TensorFlow, Kaggle, Support Vector Machines, xGBoost, Logistic Regression, Pandas
Analysis of data from past years. Feature selection and pre processing of data.
salabsa@students.kennesaw.edu
65
3/23/2018 17:54:27
Yes, the project is affiliated with KSU
Dr. Yong Shiyshi5@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
Carbon ChainJohn PriceNick MillerConnor ArmstrongAlex TurnerBrandon Gonzalez
We have developed a loyalty point program that interfaces with the Verifone Carbon platform. This system was built using blockchain technology in order to create a reward system for their network of customers. We built the system as an API to be plugged into any existing application on the point of sale device. The demonstration application was built using android studio and the proprietary sdk for the device, and the backend was developed in NodeJS with Express.js, which was hosted on AWS’ cloud platform.
BSCS
We are seeking to create a centralized loyalty point system that can be used in any business using the verifone carbon device.
We created a NodeJS web server connected to a blockchain, that interacts with the verifone carbon systems in order to facilitate transactions involving the loyalty points.
We have found that the design works thus far as a back-end implementation. At the time of this writing (3-23-18) the web server can implement transactions that are fed to it through HTTP requests to act as a stand in until the Carbon system is fully functional.
jprice58@students.kennesaw.edu
66
3/23/2018 18:16:44
Yes, the project is affiliated with KSU
Dr. Selena Heshe4@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4491YesNo
Table with power ( (ONLY IF YOU HAVE A DEMO)
Smart Alarm ClockJacob StewartKurt Floyd
An alarm clock that dynamically alters wake up times based on current weather and traffic conditions, controls room lighting, tracks user sleep patterns, and suggests that an alarm be set if a user fails to set one during a normal time.
BSCS
We wanted to try to improve the standard alarm clock, and mitigate the kinds of scenarios that have made us late to class/work in the past.
Our project is based around a Raspberry Pi B, which communicates with a remote cloud server, APIs, and local actuators. We used Python and Java for all programming.
As of now our clock has full alarm clock functionality, and pulls data from the Google Distance Matrix and Openweather APIs to change wakeup times.
kfloyd37@students.kennesaw.edu
67
3/23/2018 20:07:59
Yes, the project is affiliated with KSU
Dr. Michael Franklin
dfrank15@kennesaw.edu
Undergraduate research
NoYesEasel
eSense 2.0: Biomimetic Modeling
Derek MartinDr. Michael Franklin
We improved upon eSense 1.0, by using multiple learning layers. For our first model (eSense 1.0), we kept all of our predator observations in a single learning layer and used a sensing layer on top of that. Our results showed that this was effective, but occasionally our predator would make unexpected/unrealistic actions, such as: moving away from the prey agent when they were both near obstacles. To fix this behavior, we decided to use a multi-layered learning approach. We separated the learning into 5 layers: learning where the predators are, learning where food sources are, reacting to predators, reacting to food, and obstacle avoidance. Now, when the agents make a decision they combine the information of each layer to make the optimal choice.
BSCGDD
Can we improve our predator-prey model by using a multi-layered learning approach, for our predator and prey agents?
Unity Engine
Our results show that our new multi-layered approach creates more realistic behaviors between our predator and prey agents.
dmart133@students.kennesaw.edu
68
3/23/2018 20:15:45
Yes, the project is affiliated with KSU
Dr. Michael Franklin
dfrank15@kennesaw.edu
Undergraduate research
NoYesEasel
Peak Shaving: Optimizing HVAC Cost
Derek MartinDr. Michael Franklin
We developed a simulation, of a typical day at West Point Academy, and a predictive model to attempt to optimize the efficiency of their A/C system during peak hours of the day (when a large amount of students return to their dorm room after class or extracurricular activities). In other words, we attempted to develop a model that will preheat or pre-cool the rooms of the residents, to satisfy the residents, while also reducing the workload/cost of running the HVAC system.
BSCGDD
Can we develop a predictive model to optimize the efficiency of a HVAC system and lower the cost of running it?
Unity Engine
The results from our simulation show that our model is successful at lowering the cost compared to data provided by West Point.
dmart133@students.kennesaw.edu
69
3/23/2018 20:25:00
Yes, the project is affiliated with KSU
Dr. Mingon Kang
mkang9@kennesaw.edu
Academic courses graduate (e.g. capstones, innovative special topics projects, other course projects)
CS7267NoYesEasel
Toxic Comment Classification
Sean McGlincyShah ZafraniLinda Vu
There are negative comments happening everywhere on social media and on the Internet. This dataset consists of labeled toxic comments on wikipedia articles. Our goal is to use machine learning to predict the type of toxicity for each comment.
MSCS
Our goal is to use machine learning to predict the type of toxicity for each comment.
We will evaluate the feasibilities of different machine learning program architectures for this problem. Currently we are leaning towards RNN and SVM, but may also implement a KNN as a reference. The types of data preprocessing, and evaluation techniques: ROC curve and K-Fold cross-validation. We will also be using techniques learned in Dr. Jacob’s Text Mining course such as Natural Language Models, text classification, and probabilistic models for word frequency.

Using concepts of Inverse Document Frequency from our Text Mining course we will be able to assign a probability to each word in a comment which will serve as a heuristic for our machine learning models to use. We can also initialize our models with a bias for words that are known to be toxic (profanity, violent verbs, racial slurs). We will determine the feasibility of using Logistic Regression, followed by Support Vector Machines, and possibly a Recurrent Neural Network for classifying these comments.
Because this dataset comes from a Kaggle competition, we will be able to compare our results to those of our fellow data science peers around the world.
Data can be found at: https://www.kaggle.com/c/jigsaw-toxic-comment-classification-challenge

The data is formatted as pairs of comments and toxicity ratings (toxic, Severe_toxic, Obscene, Threat, Insult, identity_hate). The labels are not mutually exclusive, which makes our classification task rather difficult. The data is also uncensored, I have censored the sample above.
The preliminary results is to be determined.
shahzafrani@gmail.com
Thank you
70
3/23/2018 20:46:00
We are sponsored by a company called Verifone for our Senior Project
Dr. Yong Shiyshi5@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
MenYouAustin FordJackson WesselsMiguel ToledoNathan Foshee
Our project is an android application built on the Verifone Carbon 10 device. Our application is a menu management system that allows both customers and businesses to have the information of their menu at their fingertips. Our application allows businesses to create and manage their menu down to the details of the ingredients. Employees can search and filter the menu by specific ingredients and by name. This is useful for people with food sensitivities and food allergies.
BSCS
We did this project because it was offered by a leading company in a technology field. The project offered a good balance between challenge and creativity. We were given just the device and were told we could come up with any project idea we wanted. We came up with the idea for MenYou because some of our group members had food allergies or knew people with food allergies and saw this as an opportunity to solve a problem we face often.
We were loaned a Carbon 10 device by Verifone.
We have our backend Go server and our Postgresql database deployed on google cloud. Right now our application can retrieve all menu items, grab a particular menu item, create a menu item, and delete a menu item. This is all done through REST API’s. Our Android Client is currently able to make API calls to our server. We have been successfully able to deploy and APK on the Carbon 10 device.
aford26@students.kennesaw.edu
71
3/23/2018 20:48:36
Yes, the project is affiliated with KSU
Dr. Mingon Kang
mkang9@kennesaw.edu
Academic courses graduate (e.g. capstones, innovative special topics projects, other course projects)
CS 7267NoYesEasel
Diabetic Retinopathy Detection
Tejaswini MallavarapuKritika GargAmruta Mangaonkar
Using eye images of patients, we are classifying if the patient has Diabetic Retinopathy disease or not.
MSCS
Diabetic Retinopathy is leading cause of blindness. It has affected approximately 93 million of people. Currently, detecting DR is a time-consuming and manual process that requires a trained clinician to examine and evaluate digital color fundus photographs of the retina. Clinicians can identify DR by the presence of lesions associated with the vascular abnormalities caused by the disease. With color fundus photography as input, we are going to apply Machine learning techniques to build a binary classifier. This classifier will classify if the patient is diabetic or not.
Methods - CNN, SVM, LASSO, Naïve Bayes classifier
Processing
tmallava@students.kennesaw.edu
72
3/23/2018 21:03:05
Yes, the project is affiliated with KSU
Dr. Ying Xieyxie2@kennesaw.eduGraduate researchNoNoEasel
Deep Embedding Kernel
Linh LeLinhLe
In this poster, we propose a novel supervised learning method that is called Deep Embedding Kernel (DEK). DEK combines the advantages of deep learning and kernel methods in a unified framework. More specifically, DEK is a learnable kernel represented by a newly designed deep architecture. Compared with pre-defined kernels, this kernel can be explicitly trained to map data to an optimized high-level feature space where data may have favorable features toward the application. Compared with typical deep learning using SoftMax or logistic regression as the top layer, DEK is expected to be more generalizable to new data. Experimental results show that DEK has superior performance than typical machine learning methods in identity detection, classification, regression, dimension reduction, and transfer learning.
PhD in Analytics and Data Science
Design a novel supervised model integrated the strengths of deep learning and kernel methods while mitigating their individual weaknesses
Use a deep neural network to learn the kernel function usable in any kernel method
The proposed framework outperforms all other typical supervised models in all experiments.
lle12@students.kennesaw.edu
73
3/23/2018 21:50:24
Yes, the project is affiliated with KSU
Dr. Yong Shiyshi5@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CS 4850NoYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
NVidia CUDA in Bioinformatics
Brent RobinsonJustin VoorheesChip GardnerVictor Sahin
Analysis of a pairwise protein distance algorithm on an NVidia GPU using the CUDA API versus a serial implementation of the same algorithm.
BSCS
Can a serial implementation of the Phylip ProtDist program be executed on a GPU.
Phlyip ProtDist program in serial, KSU GPU Cluster with NVidia GPU, CUDA API. Methods include allocating memory and methods to be run on the GPU and testing the program against a GPU profile to determine if parallelization was successful.
Currently we have rewritten a large portion of the original serial code such that certain methods successfully run on the GPU. Optimization is being preformed to increase overall speed.
brobin79@students.kennesaw.edu
74
3/23/2018 22:07:46
Yes, the project is affiliated with KSU
Dr. Meng Han
menghan@kennesaw.edu
Undergraduate research
IT 4490NoYesEasel
Regression Analysis and Geolocation
Brittney Miller
Can meteorite landings be predicted? Upon initial review of the meteorite landing data from NASA, there appears to be a positive correlation between meteorite landings and geolocation. I will use this correlation to to analyze the dataset with regression analysis to determine the possibility of new meteorite landing site predictions.
BSIT
Can we predict meteorite landings through data analysis?
python, regression, data analytics
Upon initial review of the data, there seems to be a positive correlation between meteorite landings and geolocation. Is predicting new meteorite landing sites possible?
bmill155@students.kennesaw.edu
75
3/23/2018 23:15:06
Yes, the project is affiliated with KSU
Katina Mattis
katina.mattis@gta.ga.gov
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
IT 4983YesYesEasel
KSU Capstone 15: SCADA Networks
Allen McCroryZivalus NixonAndrew ConnellyChad WombleAshad Mahmood
An analysis of the network architecture and vulnerabilities of supervisory control and data acquisition (SCADA) systems governing the daily operation of five sectors of critical infrastructure: water, energy, communications, transportation, and emergency management. Graphical models are used to demonstrate the network layouts of autonomous control systems and how they may be attacked.
BSIT
How vulnerable are public infrastructure control systems to cyberattacks?
Research into manufacturers and vendors of SCADA systems or parts, and into how those systems are constructed. Research into historical attacks, such as the Stuxnet cyberwarfare worm, and vulnerabilities discovered by security experts or hardware manufacturers. As of the time of submission, we are exploring the options of simulating SCADA operation with software.
Preliminary results show a large range of possible attacks which can be performed on public infrastructure systems. Some are precipitated by system administrator negligence in strategically layering defense mechanisms, or by the valuation of speed and low costs over security. Others are due to the slow pace of manufacturers who are slow to update when inherent vulnerabilities are found. Securing even a single SCADA network contained in one building is an ordeal requiring a detailed understanding of how the network should communicate with the outside internet, with employee devices, with physical terminals, and even how the transmission medium used inside the network is secured.
amccrory@students.kennesaw.edu
In-progress website for project contains all deliverables currently given to project sponsor and will be updated as project continues: https://capstoneaaacz.wixsite.com/ksucapstonescada
76
3/24/2018 0:29:18
Yes, the project is affiliated with KSU
Dr. Dan Lodlo2@kennesaw.eduGraduate researchNoYesEasel
Malware Classification by ML & LBP
Jhu-Sin LuoKennesawUniversity
Malware classification is a critical part in the cybersecurity. Traditional methodologies for the malware classification typically use static analysis and dynamic analysis to identify malware. In this paper, a malware classification methodology based on its binary image and extracting local binary pattern (LBP) features is proposed. First, hexadecimal malwares are converted into RGBA scale image and reorganized into 3 by 3 grids which is mainly used to extract LBP feature. Second, the LBP is implemented on the malware images to extract features in that it is useful in pattern or texture classification. Finally, Tensorflow, a library for machine learning, is applied to classify malware images with the LBP feature. Performance comparison results among different classifiers with different image descriptors such as GIST, a spatial envelop, and the LBP demonstrate that our proposed approach outperforms others.
MSCS
The goal of this research is build up an innovative methodology to classify malware
TensorFlow, Local Binary Pattern
currently accuracy higher than 90%
jluo6@students.kennesaw.edu
77
3/24/2018 11:15:03
No, the project is not affiliated with KSU
donghyun kim
donghyun.david.kim@gmail.com
Graduate researchNoNoEasel
78
3/24/2018 13:44:41
Yes, the project is affiliated with KSU
Reza Parizi
rparizi1@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
YesYes
Table no power (ONLY IF YOU HAVE A DEMO)
M.A.I.D.A.SAndre MaxwellRobert TorresGavin SmithAdam BrittonAlex CreilsonMariel Mcneil
The purpose of our project is to use a drone in order to perform trailer yard auditing for our client, Manhattan Associates. We are to use to drone to capture the locations of trailers and their spots on a "yard" using a camera, and report the locations of those trailers to provide auditing of which trailers are on the yard and which parking spots or "docks" they reside in. This solution was asked to be deployed so that the manual labor that is required with performing manual yard audits would not longer be necessary, and the process of documenting the trailers and their locations on the yard would be streamlined.
BSSWE
Can manual auditing of trailer yards by employees be streamlined by using a drone to record trailers and their location on the yard?
Mavic Pro Drone, Mavic Pro Drone SDK, Amazon Web Services, Amazon Rekognition, Amazon Lambda, Prisma (with Graph QL), and a MYSQL database.
We will be able to perform a basic auditing of yard using the Mavic Pro drone. There are certain limitations that are more dependent upon the layout of the yard and not the functionality of the drone.
amaxwe22@students.kennesaw.edu
79
3/24/2018 14:26:55
Yes, the project is affiliated with KSU
Dr. Lodlo2@kennesaw.eduGraduate researchNoNoEasel
An new IDS using SVM and TensorFlow
Alex Federico
A network intrusion detection system (NIDS) is a tool that allows users and companies to protect their systems from malware. However, novel attacks are developed constantly, and they can create problems where new sophisticated malware can go through a computer system undetected and cause harm to the system. If a user or company is unable to train the NIDS to an updated signature within a specific amount of time, then the computer system can become vulnerable from attacks within that time period. This can be a major problem when an NIDS model needs to be trained daily, so the training time is an important feature to consider. The proposed research is to develop a new Network Intrusion Detection System by using both SVM and TensorFlow.
MSCS
Can we make a new model for an intrusion detection system that can improve the performance for training data.
Materials:
- computer installed with HPCC Systems, Python, and TensorFlow.
-network intrusions dataset containing examples of malware and normal data.

Methods:
-recording the time of the new model of the Network Intrusion Detection System by placing code at the beginning and end of the new NIDS model for start and end time, and finding the difference between the two times.
-evaluating the new model for accuracy by inputing datasets with known malware and normal data and comparing each datum to the models prediction.
No results
afederi1@students.kennesaw.edu
80
3/24/2018 14:51:47
Yes, the project is affiliated with KSU
Prof. Mike Franklin
dfrank15@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CGDD 4814YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
The Darkness WithinNadia KhanHaley Gragg
"The Darkness Within is the story of a young fox spirit who is attacked by darkness and falls from grace, waking up on Earth with no memory of its past. The Darkness now hunts the young sprite hoping to corrupt the spirit’s light and send the world into chaos. The light sprite must learn its purpose and purge the world of darkness."

The Darkness Within is a 3D platformer adventure game intended for children. The game is fantasy-based, but also teaches children simple physics as the game progresses.
BSCGDD
As a team, we wanted to make an adventure game that also incorporated simple physics lessons to begin teaching children how the world works.
Programs Used:
Unity
Maya
Substance Painter
Photoshop Elements
Audacity
(This "required" section is vague. What are you asking for? Nothing has happened yet)
nkhan25@students.kennesaw.edu
81
3/24/2018 15:01:24
Yes, the project is affiliated with KSU
Ming Yang
myang8@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
IT 4983YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
IT Alumni Database ETL System
Ricky ParksVy DuongZack DowningDesiree SmokesHang Yu
The IT department at KSU maintains an alumni database which stores graduates’ career information. In the past, alumni’s career information was manually extracted from LinkedIn.com and then loaded into a relational database which is a very labor-intensive process. In this project, we aim to create an automated extract, transform and load (ETL) system for the IT alumni database.
BSIT
The motivation for this project is to automate the extraction of information from LinkedIn, so that the IT department can retrieve and store information about Kennesaw State graduates at a faster rate.
The materials used include the Python programming language, PyCharm Community Edition IDE, and SQL Server. The general methods used include the design and implementation of the database, Python scripts used to insert and update data, and the modification of the web crawler that's used to pull information from LinkedIn.
Preliminary results include the successful extraction and insertion of data into our database used for testing.
rparks13@students.kennesaw.edu
82
3/24/2018 15:05:09
Yes, the project is affiliated with KSU
Dr. Donghyun Kim
donghyun.kim@kennesaw.edu
Graduate researchNoNoEasel
Enhancing security in smart grids
Uday Bhaskar Boyanapalli
Security vulnerabilities in SCADA communication protocol and its enhancements. DNP3 protocol in smart Grids.
MSCS
Security vulnerabilty in SCADA systems and its attacks
Deffie Helman's Keys exchange and PKI - Public key Infrastructure
Test case results on Raspberry PI
uboyanap@students.kennesaw.edu
83
3/24/2018 18:42:27
Yes, the project is affiliated with KSU
Dr. Sarah Northsnorth@kennesaw.eduGraduate researchYesNo
Table with power ( (ONLY IF YOU HAVE A DEMO)
Interaction Humanoid Robots/Humans
Joel TedoBen Tamo
Matt Alyward (High School Intern)
Project Title: Interaction between Humanoid Robots and Humans in Collaborative Environments

One of the most challenging areas of robotics research is in the cooperation between humans and machines. The main objective of this research is to investigate the interaction between humanoid robots and humans in collaborative environments. Two UXA-90s, Multipurpose Humanoid Robots (MHRs) with Scratch Builder software (RQ Robot Coding software) are used in two distinct experiments. The first experiment measures unambiguous task completion and accurateness by two MHRs working together, while the second experiment measures unambiguous task completion and accurateness by an MHR and human working together. Preliminary pilot experimentation and observations are encouraging. A comprehensive interface and experimentation are under development, and collected data are subjected to a battery of statistical analysis procedures, including ANOVA and t-test. Conclusions are derived and reported accordingly.
MSCS
Is there any significant difference between the two Humanoid Robots interacting with each other compared to a Humanoid Robot interacting with a human?
Two UXA-90s-Multipurpose Humanoid Robots (MHRs) with Scratch Builder software (RQ Robot Coding software) are used in two distinct experiments. The first experiment measures unambiguous tasks completion and accurateness between two MHRs, while the second experiment measures unambiguous tasks completion and accurateness between an MHR and human working together.
A comprehensive interface and experimentation are under development, and collected data are subjected to a battery of statistical analysis procedures, including ANOVA and t-test. Conclusions are derived and reported accordingly.
kjol@students.kennesaw.edu
84
3/24/2018 19:09:59
Yes, the project is affiliated with KSU
Dr Dan Lodlo2@kennesaw.edu
Undergraduate research
HON 3002YesYesEasel
MS Office Macro Malware Detection
Ruth Bearden
Macro malware hidden in Microsoft Office files still poses a threat to cyber security two decades after its emergence. This research seeks to increase macro malware detection accuracy by using machine learning techniques to improve common approaches to the problem.
BSCS
Can we improve current macro malware detection methods by using machine learning? Macro malware is malicious code hidden inside Microsoft Office files distributed as email attachments. It has plagued corporations for decades since it emerged in the late 20th century, and it is still relatively immune to antivirus software. Commonly used approaches to detecting macro malware involve scanning MS Office files for keywords indicating malicious code, but clever hackers find new ways to avoid these keywords and the red flags they raise. Our method uses machine learning techniques to get more complete statistical information from a document than these common methods do and use it to detect macro malware with better accuracy.
Methods - Our method for detecting MS Office macro malware is to take a set of sample MS Office files (Word and Excel - some containing macro malware), "learn" from these, and use what was learned to make a predictive model used to determine whether or not any given file is likely to contain macro malware. To implement this, we converted the problem to a machine learning "binary classification" problem where each file was classified as either "malicious" (containing macro malware) or "benign" (harmless). We used each file's macro code and classification label to train a classifier and tested our methods on a randomized subset of the sample we gathered. The proof-of-concept stage of this project used the K-Nearest Neighbors machine learning classifier - one which is simple to understand and implement. At present we are working to test and compare a variety of classifier algorithms on our sample set including Naive Bayes, Neural Network, and SVM.

Materials - In the proof-of-concept stage, we implemented our methods on a single computer using Java and Python and used this platform to process 158 samples. Since we have recently doubled our sample set size and because the algorithms we wish to evaluate on our data require a lot of processing resources, we are currently working to implement our methods using Python and ECL on the HPCC Systems platform which facilitates parallel processing of large amounts of data on a multi-node computer network.
At the proof-of-concept level, using the K-Nearest Neighbors algorithm and representing files in our sample set by a compiled version of their macro code, we identified files containing macro malware with 95% accuracy.
rbeard13@students.kennesaw.edu
In the current stage of this research we are working to compare the effectiveness of different machine learning algorithms and feature selection methods in detecting macro malware. This stage is still in progress but we anticipate having the full results and comparison come C-Day.
85
3/24/2018 19:36:46
Yes, the project is affiliated with KSU
Dr. Reza Parizi
rparizi1@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
SWE 4724YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
Roommates40Plus.com
John LandsmanFerit AkaybicenDaniel HanleyDaryan SankarAndres RodriguezForrest Cain
Roommates40Plus.com is a dedicated website for older US citizens to find potential roommates. Generic forum websites such as Craigslist are looked down on by older citizens as a dangerous place for criminals and con artists. Roommates40Plus.com offers a safe and simple place for older generations to find like minded individuals to meet and potentially live with. This project was presented to us by Ms. Susan Asher, a Marietta local and owner of Roommates40Plus.com.
BSSWE
As the Baby Boomer and Gen X population approach retirement, an increasing number of older US citizens are struggling financially. There are a myriad of reasons why these citizens are struggling; illness/injury, termination from a job, and the death of a spouse are just a few of those reasons. So what can a Baby Boomer or Gen X do to alleviate the cost of living? Get a roommate! But where? Older citizens are particularly wary of utilizing websites such as Craigslist, so they need a dedicated platform to help them look and Roommates40Plus.com is that platform. Through this service, users can create an account and list their likes, dislikes, pet peeves, and other deal breakers when it comes to finding the right roommate. With that information along with their age and location, users are matched with potential roommates.
Materials

+ Front End
- HTML5 (markup language)
- CSS3 (style sheet language)
- JavaScript with React library (programming language)

+ Back End
- GO (programming language)
- MySQL (database management)
- JavaScript

+ Services
- GitHub (code repository)
- Google Drive (document sharing)
- Slack (communication and sharing)
- GroupMe (communication)
- Blue Host (web domain registrar)
- Amazon Web Services (database and web hosting)

Methods

For this project we decided to utilize the agile development model. The reason for this is because our team is small and there is currently no existing platform for us to work from.
As of today (March 24, 2018) we have a React webpage up on a mock domain and plan to have the website up on AWS by March 31st.
jlandsma@kennesaw.edu
86
3/24/2018 21:08:36
Yes, the project is affiliated with KSU
Dr Ying Xieyxie2@kennesaw.eduGraduate researchYesYesEasel
Monitoring animals in wilderness
Joel Kamdem Teto
The purpose of this research is first to be able to localize animals in wilderness using live camera feeds. Then, using the existing techniques (Convolutional Neural Network) detect each animal, count for each species and identify the behavior at the moment. The behavior here is considered as either standing, not standing, resting or not, eating, baby. Nonetheless, a recent paper from Geoffrey Hinton, have proven some shortcomings of the state-of-the-art convolutional neural network to detect objects. Indeed, he proposed a new model of network called the Capsule network that is supposed to overcome the shortcomings of the CNN. The main objective of this research is to be able to achieve this task using CNN and Capsule Network.
MSCS
Is Capsule Network more accurate than the state of the art CNN in detecting lively objects like animals? Can Capsule Network improve classification in images? (defining behavior of animals)
Using deep learning, i will subsequently do detection of animals on different settings, dark, long distance, day light, etc... then i will focus on counting the specimens and finally focus on the attributes.

References:
Hinton, G. E., Osindero, S., and Teh, Y. (2006). A fast learning algorithm for deep belief nets. Neural Compu-tation, 18, 1527

Norouzzadeh, MS, Nguyen, A, Kosmala, M, Swanson, A, Palmer, M, Packer, C, & Clune, J 2017, 'Automatically identifying, counting, and describing wild animals in camera-trap images with deep learning'.
Shahroudnejad, A, Mohammadi, A, & Plataniotis, KN 2018, 'Improved Explainability of Capsule Networks: Relevance Path by Agreement'.
working with Capsule network on digits detection as a test.
downloading a dataset of 3 millions images having different settings to apply the capsule network and the cNN
kjol@students.kennesaw.edu
87
3/24/2018 22:42:01
Yes, the project is affiliated with KSU
Dr. Selina Heshe4@kennesaw.edu
Undergraduate research
CS 4491NoYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
Ride Experience Nnamdi NwajaguAdnan RasheidNnamdiNwajagu
A small device that will be placed on the bottom of board and monitor speed, acceleration, altitude and orientation to recognize tricks and various trip ride statistics.
BAACS
Creating a device that will let a user monitor their performance on a skateboard.
Raspberry Pi, Skateboard, US5881 Hall Effect Sensor, MPU5930
A device that can record statistics during a skateboard ride while enduring the stresses of being underneath a skateboard is viable.
nnamdi.nwajagu@gmail.com
88
3/24/2018 23:25:50
Yes, the project is affiliated with KSU
Dr. Michael Franklin
dfrank15@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CGDD 4814NoYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
SquadsDerek Martin
Can you command your squad and lead them to victory, in this first person tactical shooter? Choose your team strategy and give commands on the fly to get a drop on your opponents.
BSCGDDCapstone ProjectUnity Engine
I have developed a playable demo.
dmart133@students.kennesaw
89
3/24/2018 23:43:21
Yes, the project is affiliated with KSU
Dr. Ken Hoganson
khoganso@kennesaw.edu
Graduate researchCS 7999NoNoEasel
Music Generation w/ Neural Networks
Shah Zafrani
Researchers have shown compelling results in using layered recurrent neural networks trained on classical music for raw audio music generation. In this project we explore variations on existing architectures that can handle genres with more varying melodic structures such as Math Rock and Progressive Metal.
MSCS
Can Recurrent Neural Networks generate novel samples of Math Rock and/or Progressive Metal.
Theano, Python, SampleRNN
Inconclusive
szafran1@students.kennesaw.edu
Work in progress. Exciting results to come. Will make a great cday poster.
90
3/24/2018 23:55:14
Yes, the project is affiliated with KSU
Reza Parizi
rparizi1@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
SWE 4724NoYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
HealthQBenny VillegasEddie MatosAustin HarrisZikomo BullockStuart TresslarBrian Deguzis
The U.S. Centers for Disease Control and Prevention (CDC) has recognized that a need exists for a free, fast-to-learn, easy-to-use software solution that is designed around epidemiologists and their public health initiatives. HealthQ rises to the challenge with an application using Google’s SDK and it’s virtual voice intelligence, Google Assistant. The purpose of this project is to improve the quality of research the CDC is conducting while alleviating the hardships of researchers on the field. HealthQ's goal for their product is to help the CDC in their efforts to prevent disease, injury, and disability, promote health and well-being, and prepare for and respond to disasters. HealthQ believes they can aid in this efforts by analyzing data and drawing conclusions based on user prompted questions. HealthQ believes anyone should be able to conversant with their data.
BSSWE
Our motivation for this project to is jump start the CDC with incorporating virtual assistant technology in their research. Whether it's to help an epidemiologist get acquainted with their data sets, free up researcher's and medical examiner's hands on the field, or give the visually impaired a chance to interact with data, HealthQ believes everyone has the right to vocally interact with data.
Virtual Assistant Technology, Google, Google Assistant, Google Mini
Reference data from select CDC databases and asking it simple questions about a data set sample.
villegas.benny@gmail.com
Our demo requires an unrealistic level of quiet because we'll be talking to a Google Mini.
It may work, however, depending on where we are.

Thank you!
91
3/25/2018 0:57:49
Yes, the project is affiliated with KSU
Walter Tong
walter.tong@gta.ga.gov
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
IT4983NoYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
Cyber Actor Profiles Ramiz RamicSabrina BuckholdtKevin TillackHao PhamNick Kennelly
Research of cyber actor groups from: Eastern Europe, Asia, and Middle East. Develop profiles of each group's activity to include motivations, techniques, tools, and procedures and favorite targets of attack. Video demonstration of a group's typical attack against a favorite target.
BSIT
Based on previous attacks, risks and weaknesses of CIKR (Critical Infrastructure and Key Resources), determine which cyber actor group would attack CIKR in Georgia
Research Document, Kali, VM construction, Spear Phishing. Remote Access Trojans
Dragonfly group uses remote access tools to execute attacks against energy sectors and create a back door on computers. APT 33 uses spear phishing emails to execute attacks against the aviation industry. APT 10 uses remote access tools to execute attacks against health care and defense critical infrastructures. Research and one multimedia attack demonstration for each cyber actor group completed.
rramic@students.kennesaw.edu
92
3/25/2018 9:22:04
No, the project is not affiliated with KSU
Stephanie Gambro
stephanie.gambro@adp.com
Service (e.g. internship, capstone project, student club)
CSE4983NoYesEasel
93
3/25/2018 9:44:12
Yes, the project is affiliated with KSU
Dr. Mingon Kang
mkang9@kennesaw.edu
Graduate researchNoYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
R-CNN for Gender Identification
Rehnuma AfrinDr. Mingon KangRehnumaAfrin
R-CNN (Region based Convolutional Neural Network ) is very effective for object detection and localization. It takes in an image, and correctly identify the objects in the image via a bounding box. In this project, R-CNN is used for Gender detection and classification.
MSCS
I have been working with R-CNN for last few months. My supervisor motivated me to apply R-CNN for gender detection and localization.
Materials: Alexnet, Caffe, Support Vector Machine, Selective Search Algorithm, Non-Maximum Suppression.
After training the SVM using the features from CNN, next task is to input test images. Selective Search will be applied for object detection followed by forward propagate it to CNN and Non Maximum Suppression. And finally getting the labels along with bounding box.
Worked with another dataset.
rafrin@students.kennesaw.edu
Not 100% sure whether I can finish the DEMO part. Just to be in safe side, in equipment requirements : I applied for Table with No power.
94
3/25/2018 10:17:52
Yes, the project is affiliated with KSU
Dr. Reza Parizi
rparizi1@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
SWE 4724YesYes
Table with power ( (ONLY IF YOU HAVE A DEMO)
Web Therapy Application
Nygel JonesTrey Hill Geovanni AmaizoBrad CropseyTevi DrackeyPhuc Nguyen
Our project was to create a website that allowed patients with autism to find, view, and schedule an appointment with therapists in their area.
BSSWE
What would be a way to make therapy more accessible to those that need it.
Poster board, computer, internet access

The website was created
njones99@students.kennesaw.edu
95
3/25/2018 11:30:50
Yes, the project is affiliated with KSU
Dr Mingon Kang
mkang9@kennesaw.edu
Graduate researchCS7999NoYesEasel
Section Extraction from pdf
sai chandra kosaraju
Extraction of sections from unstructured data (example PDF) . extraction of features from the unstructured data, using this features extraction of important information from the documents
MSCS
conversation of OCR to text blob is challenging task extraction of OCR data in customized formats will help the industries to build their own Ontologies
Pdf miner modules Machine learning techniques like fuzzy logics
we converted PDFs to JSON format with easy user interface
skosara1@students.kennesaw.edu
96
3/25/2018 11:50:11
Yes, the project is affiliated with KSU
Dr Mingon Kang
mkang9@kennesaw.edu
Academic courses graduate (e.g. capstones, innovative special topics projects, other course projects)
CS7267NoYesEasel
Demographic Segmentation
Sai chandra Kosaraju
Uday Bhaskar boyanpally
we implement machine learning techniques like deep neural network,LASSO ,logistic regression to analysis the features which effect the customers leaving leaving the bank
MSCS
Analysis of risk factors effecting the company
LASSO,deep neural networks ,logistic regression
we got the probability for the customers data
skosara1@students.kennesaw.edu
97
3/25/2018 13:41:53
Yes, the project is affiliated with KSU
Dr. Dan Lodlo2@kennesaw.eduGraduate researchCS 7999YesYesEasel
Predicting Amazon Best-Sellers
Michael Kranzlein
A multiple classifier system of neural networks built on Amazon product descriptions and images for predicting best sellers across a variety of product categories.
MSCS
Are sales ranks related to product descriptions and product images? Can sales rank be predicted from just these data points?
Convolutional Neural Networks and Recurrent Neural Networks via TensorFlow
70%+ accuracy in individual models
mkranzle@students.kennesaw.edu
98
3/25/2018 13:46:16
Yes, the project is affiliated with KSU
Dawn Tatum
dtatum7@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
IT 4983NoYesEasel
CCSE Internship video blog
Juan Castro Isaura Romero Joe BlackAlexander Lam Evgeniya Koganitskaya
We are exploring the possibility of creating a website for internship students to post video blogs/presentation about their internship experiences.
BSIT
To create a KSU website to facilitate scholarship awareness targeted to students through a video blog.
Wordpress, YouTube v3 data api, html, css and javascript
Website is on the final stage of development, Professor Tatum will be able to upload the videos of her choice into her admin page and then approve them to be seen in the public website.
jcastro9@students.kennesaw.edu
99
3/25/2018 13:57:53
Yes, the project is affiliated with KSU
Dawn Tatum
dtatum7@kennesaw.edu
Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)
CSE 4983NoYesEasel
Internship at AMB Group
Juan Castro
The internship is for a Desktop Support Assistant, the responsibilities are to help clients with software and hardware issues. We have to set up and image computers for new employees. We have to monitor a help desk and resolve any tickets assigned to our department. Tools we are using include: Active Directory, ServiceNow, Office 365, and Maas360
BSIT
Apply knowledge through industry experience
n/a
I have learned how to use Active Directory, servicenow and office 365. I've helped the company close 15-20 trouble tickets on average on a weekly basis.
jcastro9@students.kennesaw.edu
100
3/25/2018 14:13:24
Yes, the project is affiliated with KSU
Dr. Chih-Cheng Hung
chung1@kennesaw.edu
Graduate researchCS7999NoYesEasel
Algorithms to Improve MRI Scanning
Srivarna Settisara Janney
Dr. Sumit Chakravarty
Applying three new algorithms in Compressed Sensing to Medical images (MRI) such as Brain, Chest, Shoulder, and Heart type of image data sets, offers significant scan time reductions, which benefits patients, at the same time retains quality images for accurate diagnoses by radiologist.
MSCS
How to reduce time for patient who go through a painful ordeal of MRI scan? How accurately we can reconstruct MRI data after scanning?
Brain, Chest, Shoulder and heart type of dataset are used. Matlab is the platform of implementation.
We developed 3 new algorithms namely Weighted Threshold Wavelet Tree Structure (WTWTS), Directional Wavelet Tree Structure (DWTS) and Weighted Directional Wavelet Tree Structure (WDWTS).
Final results are amazing. Image reconstructed from my algorithms produced 20% better image quality and at 5-10% reduction in the time taken to run the algorithm when compared to current state of art reference paper results.
ssettisa@students.kennesaw.edu
Loading...
Main menu