Facial Emotion Recognition with Convolutional Neural Networks

Dzung Pham

Final project for CS 373 Artificial Intelligence - Williams College

Predict facial emotions expressed in real-time with your webcam. The 7 emotions are: angry, disgust, sad, happy, fearful, surprise, neutral.

Test accuracy: 65.33%. Validation accuracy: 64.25%

Sample images:

Installation

See readme.txt for detailed instructions on how to set up and experiment with my code.

Model

You can download my CNN model at: https://drive.google.com/open?id=1Y_BnRydcYhWiGbEaw74aRXswbn9qAKCP

Dataset

To get dataset, create a Kaggle account and request data from: https://www.kaggle.com/c/challenges-in-representation-learning-facial-expression-recognition-challenge/data

Report

See my report for more details and references: https://github.com/vietdzung/facial_emotion_recognition_cnn/blob/master/report/final_report.pdf

Special thanks

The work of these people is my source of inspiration: https://github.com/isseu/emotion-recognition-neural-networks