Edge Computing Face Detection
Project Overview
This project implements face-detection on a low cost Raspberry Pi computer with CCD camera. A CCD camera captures images in real time. Face detection and recognition are made using OpenCV image library. Since the system is low cost and low power it can be installed cheaply and easily in remote locations. This allows many practical applications.
Typical Applications
The are many uses for a system able to recognise someone that has been seen before. Examples include:
- understanding repeat customer behaviour in retail environments
- Identify and flag banned individuals in nightclub settings
- Enforce protection orders to automatically flag when banned individuals approach restricted areas
Implementation
OpenCV library can be trained on open-source face datasets. This will allow face morphology to be converted into a biometric data point. New faces can be allocated and recognised once trained on the source data.
Face detection algorithms are able to separate individual faces within a frame and apply the face detection strategy to each face observed.
The point is to train the system to recognise individuals it has seen before and allocate a unique UUID to the face. Actions taken by the system then depend on configuration.
