Skip to content

Realtime Facial Emotion Recognition Application

Notifications You must be signed in to change notification settings

Ap00rvSaxena/FERS

Repository files navigation

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.

Dataset

Available at kaggel: FER2013

Online Model Available at Google Colab

Available at Colab: FERSystem

Demo

Neutral Face

Sad Face

Angry Face

Surprised Face Happy Face