Face Detection, Recognition and Expression Recognition

The repo contains the code for running the All-in-one face detection/recognition + expression recognition

Installation

Requires several libraries to get the demo running.

1. Caffe + PyCaffe Installation
2. Tensorflow for Face expression Installation
3. OpenCV
4. Currently the code supports NVIDIA GPUs. For CPU runs we may need to modify the code.
5. Requires Python 2.7

Usage

1. Download the model files from here
2. Untar the contents and place the content in the models directory
3. Place the images that you want to be identified in the data directory. Then run

 python utils/generate_train_features.py 

4. Next, run

 python demo.py 

A window will open and will start detecting the faces

5. Experimental feature. For detecting the facial expressions, modify utils/MTCNN.py and change EXPRESSION_DETECTION_ENABLED = True and restart the demo.py

TODO: Expression recognition is buggy and the trained model is inaccurate. Need a better model for improved accuracy.

Contributing

For major changes, please open an issue first to discuss what you would like to change.