LEAR - Local Evidence Aggregation for Regression Based Facial Point Detection

IMPORTANT: This code is compiled for 64 bits Windows


LEAR detects 20 fiducial facial points. For a detailed description of the algorithm, please refer to:

B. Martinez, M. F. Valstar, X. Binefa, M. Pantic. Local Evidence Aggregation for Regression Based Facial Point Detection. IEEE Transactions on Pattern Analysis and Machine Intelligence. 35(5): pp. 1149 - 1163, 2013.


If you don't have the matlab runtime component installed, or have a different Matlab version than 2012a, then it might be necessary to install the 2012b Matlab Component Runtime. It can be downloaded from the same page as the executable, and it will not cause any crash with other current Matlab versions installed in your computer



LEAR can be run from the command shell or directly from Matlab. From matlab you can use the function "system" to simulate a call from the command shell. 

The input can either be a directory with images:

LEAR.exe directory_path extension

or an image path

LEAR.exe image_path


IN:
- "extension" is OPTIONAL. If not introduced and "directory_path" is a folder, then it will search for all the .png files

OUT:
The program creates the following files in the same directory as the image:
- an xml with the detected point information
- a .mat with a matlab data file containing the detected points
- a .jpg file with the original image and the detected points marked on top

Please refer to the article to see the definition of point locations and the names used in the XML file.


If you encounter any problem, please contact us and give us all the information needed to reproduce the problem