RealTime Facial Emotion Recognition System
The goal of the project is to identify 7 key human emotions: HAPPY, NEUTRAL, SAD, SURPRISE, ANGER, DISGUST and FEAR, the project uses FER2013 dataset to train a classification Convolutional neural networks (CNN) model which takes a grayscale images of faces and predicts the most probable emotion in the output layer. The model traning was done on google colab and it took near about 6.5 hours to fully train this model at 200 epochs trained model was then used as the baseline to classify a facial emotion in the real-time application which was implemented using OpenCV, we extract the detected face from image frames in a real-time video then offer it to the model to predict and print the emotion in real-time.
Available at kaggel: FER2013
Online Model Available at Google Colab
Available at Colab: FERSystem