eigenfaces

This project uses a set of black and white images of people's faces to generate a set of eigenfaces, which represent the principal axes of statistical variance in the image set.

The example can be run with main.py

This can then be used to classify the faces for example as smiling or not smiling.

Here is the average face from the training set:

Here we see the first 16 computed smiling eigenfaces and non-smiling eigenfaces