Face recognition on the Pi Zero

Run one-shot face recognition on your Raspberry Pi Zero with the help of the Movidius Neural Compute Stick.

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes.

Hardware Prerequisites

• 1 NCS device
• A webcam supported by the Pi
• Button (optional)

Software Prerequisites

First you need to install requirements for the following software,

• dlib and face_recognition (https://gist.github.com/ageitgey/1ac8dbe8572f3f533df6269dab35df65)
• NCS API (https://movidius.github.io/blog/ncs-apps-on-rpi/)

Details

[Button trigger] -> [Image capture] -> [Face detection] -> [Face recognition] -> [Face classification]

A KNN classfier needs to be trained first on a set of known faces before we can start classifying. Update the known_faces directory with folders of people's names and their respective face pictures in these folders.
Then, clone the repo into your Pi.

git clone https://github.com/prampey/Pi_Face_recognition.git
cd Pi_Face_recognition

Now train the classifier:

python3 train_classifier.py

The classifer model will be generated in the model folder. Now you can run video_face_matcher.py.

python3 video_face_matcher.py