Batch-Incremental and Robust Principle Component Analysis

摘要

Incremental principle component analysis (IPCA) has been of great interest in computer vision and machine learning. In this paper, we introduce a new IPCA method, and extend it to robust PCA. The proposed method can keep an accurate track of the mean of the data, and can deal with a set of new observed data in batch each time in subspace updating. Furthermore, a weighting function can be simply incorporated in the proposed method. The performance of our method is illustrated in two experiments on face modeling and background modeling.