Investigating the boosting framework for face recognition
Bas Boom, Robin van Rootseler, Raymond N. J. Veldhuis · University of Twente Research Information · 2007
The boosting framework has shown good performance in face recognition. By combining a set of features with Adaboost, a similarity function is developed which determines if a pair of face images belongs to the same person or not. Recently, many features have been used in combination with Adaboost, achieving good results on the FERET database. In this paper we compare the results of several features on the same database and discuss our solutions on some of the open issues in this method. We compare the boosting framework with some standard algorithms and test the boosting algorithm under difficult circumstances, like illumination and registration noise.