Abstract
This paper proposes a method that can be used to detect non-occluded and occluded faces by using face parts, such as eyes and lips, as well as whole faces. The proposed method involves AdaBoost-based classifiers for whole faces and individual face-parts trained on non-occluded face sample sets. Whole faces and their parts are classified individually and the final decision is made by combining the outputs from all the classifiers. We used linear discriminant analysis and a decision tree to combine the outputs. The experimental results revealed that the proposed method was extremely effective in detecting both non-occluded and occluded faces.
