Welcome to MPLab CERT Computer Expression Recognition Technology. 

To run the software on example images, open MainScript.m and read the
comments. We suggest running the commands one at a time by pasting
them into a matlab window.  This way you can see what each component
does, and what its input and output is.

This release (CERT_1.0.0) uses
Locator: 2003 Face_and_Eye_Finder 
Classifiers: linear SVMs on Gabor representations
Actions: trained on 8 Facial Actions: 1 2 4 5 10 12 14 20
Training data: stills from Ekman-Hager Cohn-Kanade posed expressions

To run the auto facs coder you need the path to a directory where the
images are stored (Eg testImages1). Recognition output and eye detector output
are saved in a mat_file (Eg testImages1out.mat), as well as
tab-delimited text (Eg testImages1out.txt). A video demo of the system
output is also created and saved. (Eg testImages1out.avi).