Facial age classification using subpattern-based approaches

Abstract

In this paper subpattern-based approaches are used to solve the age classification problem on facial images. Subpattern-based approaches named Local Binary Patterns (LBP), subpattern-based Principal Component Analysis (spPCA) and modular Principal Component Analysis (mPCA) are examined to demonstrate the age classification performance on female and male facial images of human beings with several parameter settings. Classification of age intervals are conducted separately on female and male facial images since the aging process for female and male is different for human beings in real life. Subpattern-based LBP, spPCA and mPCA are used for feature extraction on different datasets selected from FG-NET and MORPH databases. Experimental results demonstrate the superiority of subpattern-based LBP over spPCA and mPCA techniques. Age classification performance using these three subpattern-based techniques with different parameter settings on the selected datasets is also presented.