Product-Based Capstone
LeukoVision
Team ID : C242-PS314
Meet Our Team
Machine Learning
Cloud Computing
Mobile Development
Muhammad Thayyib
Refan
Dinufalah
Nazulasari
Aes Saputra
Madona Sintiyawati
Khairunnisa
Rizky
Fauzi
Machine Learning Team
Muhammad Thayyib
Universitas Negeri Padang
Informatika
Refan Dinufalah
Universitas Siliwangi
Sistem Informasi
Nazulasari
Universitas Raharja
Sistem Informasi
Cloud Computing Team
Aes Saputra
Universitas Bina Bangsa
Sistem Informasi
Madona Sintiya
Universitas Bina Bangsa
Sistem Informasi
Mobile Development Team
Khairunnisa
Universitas Telkom
Sistem Informasi
Rizky Fauzi
Politeknik IDN Bogor
Ilmu Komputer
Background
LeukoVision originated from the awareness of the high number of leukemia cases in Indonesia, particularly among children and young adults, which has become a significant concern within the medical community. Data indicates that leukemia is one of the most common types of blood cancer and requires early detection to improve patient survival rates. Aiming to support a faster and more accurate diagnosis process, our team developed a technology-based application that integrates TensorFlow for object detection and image classification to provide efficient and reliable leukemia cell analysis.
Reason
Benign
malignant
The reason behind developing LeukoVision is to address the pressing need for early detection of leukemia cells, which is crucial for improving patient outcomes. With leukemia being one of the most prevalent types of blood cancer, particularly affecting vulnerable age groups, our team recognized the potential of leveraging technology to enhance the speed and accuracy of diagnosis, ultimately supporting medical professionals in delivering better healthcare.
Blood cell Cancer [ALL]
Leukemia
link : https://www.kaggle.com/datasets/mohammadamireshraghi/blood-cell-cancer-all-4class
Dataset
Implementation and Reasoning
1. Utilizing TensorFlow to detect and classify leukemia cells from microscopic images. This approach is designed to improve the speed and accuracy of cell analysis, supporting medical personnel in making diagnostic decisions.
2. The application is developed using Kotlin in Android Studio, with a focus on an intuitive user interface. The goal is to provide medical personnel with quick and efficient access to leukemia cell analysis results.
3. MySQL is used as the database for storing and managing medical data securely. The backend is integrated with a custom API to ensure seamless data retrieval and storage for medical personnel, ensuring data reliability for further analysis.
Training and Validation
Tools
Api Architecture
Api Documentation
Demo Model Analyze
Demo Model Analyze
System Architecture
Demo Mobile Development
Demo Model Analyze
Demo Mobile Development
Summerize
We are responsible for designing and developing an intuitive and user-friendly Android application using Kotlin. We ensure the UI/UX design aligns with the needs of medical personnel for quick and easy access to leukemia cell analysis results. We also implement data caching mechanisms using local storage to improve performance and handle any errors during data retrieval.
We are responsible for building and training a machine learning model using TensorFlow to detect and classify leukemia cells from microscopic images. Our tasks include creating a dataset, reducing model overfitting, and deploying the trained model on a local server to ensure accurate analysis results.
We are responsible for creating and managing APIs for secure data communication between the mobile app and the database. Using MySQL, we ensure reliable storage of medical data and implement robust authentication mechanisms. Additionally, we deploy the backend and ML model on a secure server, ensuring seamless integration and cost management.
We hope this application can continue to evolve to meet the broader needs of medical personnel. In the future, additional features such as medical action recommendations based on detection results, patient statistical analysis, and integration with other medical devices can be added to enhance its benefits. Moreover, we aim for this application to be accessible in various regions, especially remote areas, to support early leukemia detection quickly and accurately, ultimately improving patient survival rates. Data security and ease of use will remain our top priorities in the continuous development of this application.