Adaptive Weighted Nearest Feature Space Analysis and Its Application to Feature Extraction

摘要

In this paper, a new feature extraction algorithm named Adaptive Weighted Nearest Feature Space Analysis (AWNFSA) is proposed. AWNFSA is a Nearest Feature Space (NFS) based subspace learning approach. In Discriminant Nearest Feature Space Analysis (DNFSA) algorithm based on NFS, it may lead the result into misclassification when the between class scatter is very big or within class scatter is very small. Different from DNFSA, AWNFSA evaluates the effect of two scatter for classification through choosing their weights adaptively. The proposed AWNFSA is applied to image classification on ORL face Database. The experimental results demonstrate the efficiency of the proposed AWNFSA.