FER - Facial Expression Recognition

This work is to demonstrate the below problem: https://www.kaggle.com/c/challenges-in-representation-learning-facial-expression-recognition-challenge

A real time face detector and emotion classifier is built using Convolution Neural Network and OpenCV. The CNN model is tuned for fine performance even on a low end device.

Instructions

Follow the guided tutorial for neural network training.

Files Structure:

FER_CNN.ipynb - Tutorial to train the CNN
FER.py - Uses the pre-trained model to give inferences
model.json - Neural network architecture
weights.h5 - Trained model weights

Installation

Using Python virtual environment will be advisable.

For model prediction

pip install -r requirements.txt OR pip install opencv-python pip install tensorflow (Note here we are installing tensorflow-cpu) pip install keras