This is a face recognition demo written in Python using OpenCV. It uses a cascade classifier for local binary patterns. It can recognize human faces in real-time video streams, e.g. from a webcam.
You can clone the git repository via git clone https://github.com/czielinski/facerecognition.git. To run the code you need the numpy and the opencv libraries. On a Debian based system you can install them via apt-get install python-numpy python-opencv.
To run the program change to the facerecognition/ directory and start it with python facerecognition.py. It will then record a video stream from your webcam and do face recognition in real-time as shown here:
I trained the cascade classifier in this demo using the tools provided in this repository. A dataset of cropped faces which can be used for training can be found in the LFWcrop Face Dataset, which is a cropped version of the Labeled Faces in the Wild database.