Ping-Han Lee, Kuang-Yu Chang, Wen-Yan Chang , Shen-Chi Chen, Chih-Wei Lin, Yu-Shan Lin
We propose the Facial Trait Code (FTC) to encode human facial images. A given face can be encoded at some prescribed facial traits to render an n-ary facial trait code with each symbol in its codeword corresponding to the closest Distinctive Trait Patterns (DTP). In order to handle the most rigorous face recognition scenario in which only one facial image per individual is available for enrollment and face variations caused by illumination, expression, pose or misalignment. We also propose the Probabilistic Facial Trait Code (PFTC), with a novel encoding scheme and a probabilistic codeword distance measure. We also propose the Pattern-Specific Subspace Learning (PSSL) scheme that encodes and recognizes faces robustly under aforementioned variations.