Comparative Graph Model
Unsupervised facial recognition technique via similarity based data cleaning.
This directory contains code for an implementation of the Comparative Graph Model algorithm. The main algorithm is found under "NEW_TEST.vb" and other .vb files are supporting files that are needed for the algorithm. The algorithm in this code is set up for a training set of ten images, however it is easily extended to be N images.
The machine learning algorithm used in this experiment is a single layer perceptron algorithm and can be found under "FaceComparer.vb".