LOCAL BAGGING AND ITS APPLICATION ON FACE RECOGNITION

Transactions of Nanjing University of Aeronautics & Astronautics

Archive > Volume Issue 3 > 2010(3):255-260 Prev Next

LOCAL BAGGING AND ITS APPLICATION ON FACE RECOGNITION

Article
Figures
Metrics
Preview PDF
Reference
Related
Cited by
Materials

Abstract:

Bagging is not quite suitable for stable classifiers such as nearest n eighbor classifiers due to the lack of diversity and it is difficult to be direc tly applied to face recognition as well due to the small sample size (SSS) prope rty of face recognition. To solve the two problems, local Bagging (LBagging) i s proposed to simultaneously make Bagging apply to both nearest neighbor classifi ers and face recognition. The major difference between LBagging and Bagging is that LBagging performs the bootstrap sampling on each local region partitioned from the original face image rather than the whole face image. Since the dimensi onality of local region is usually far less than the number of samples and the c omponent classifiers are constructed just in different local regions, LBagging deals with SSS problem and generates more diverse component c lassifiers. Experimental results on four standard face image databases (AR, Yale , ORL and Yale B) indicate that the proposed LBagging method is effective and robust to llumination, occlusion and slight pose variation.

Keywords:

Project Supported:

Clc Number: