Learning Facial Pattern Classifier by AdaBoost

Bo Wu


Abstract

I will present my previous work in Tsinghua University. The topic is learning facial pattern classifier by AdaBoost. This work is based on the AdaBoost based face detection method proposed by viola & Jones, 2001. Our contributions are: 1) Using real AdaBoost instead of discrete AdaBoost; 2) Learning LUT type classifiers instead of threshold type ones; 3) Extending the frotal FD system to rotation invariant, multi-view, and omni-directional situations; and 4) Applying real AdaBoost on demographic classification problems. Here is some links to the relative publications:


Maintained by Philippos Mordohai