1 of 4

OUR FINAL PROJECT

Facial Expression Recognition (FER) Tool for Nonverbal Communication

Minseo Choi

Andrea Diaz

4 August 2023

2 of 4

PURPOSE

An algorithm that detects 5 facial expressions and outputs the corresponding text/emoji.

Convolutional Neural Networks

Artificial Intelligence algorithm that specializes in image recognition and processing by applying filters to detect patterns and features.

CNNs

Nonverbal Communication

ASD

"Past research suggests that individuals with autism spectrum disorder (ASD) present difficulties to decode facial expressions."

- National Library of

Medicine

Our purpose –

We wanted our final project to be something beneficial to a community or group of people, and this is where we got our idea. After a period of research, we learned that many individuals with autism spectrum disorder (ASD) have a harder time reading facial expressions compared to neurotypical people, and so we decided to create a program that would aid in reading people’s emotions to enhance nonverbal communication. Our algorithm is able to distinguish five different expressions: angry, happy, sad, surprised, and neutral.

A: The system is based on Convolutional Neural Networks, which in simple terms is an Artificial Intelligence model that specializes in image recognition and processing by applying filters to detect patterns and features.. We used a dataset with about 29,000 classified expressions, and then split the data so that the program is trained on 75% of it and tested on the remaining 25%

3 of 4

DEMO & RESULTS

4 of 4

Thank you!

TO THE FACULTY AND MENTORS