Combination of PCA and Wavelet Transforms for Face Recognition on 2.5D Images
Chi-Fa Chen, Yu‐Shan Tseng, Chia‐Yen Chen · 2003
This work presents a method by which increased accuracy in face recognition using the Principal Components Analysis (PCA) on wavelet transforms is achieved by using 2.5D depth maps as the source of facial features. Comparable or better results are yielded in less processing time under tested conditions. A variety of classifiers are employed, such as the nearest centre (NC), nearest feature line (NFL) and linear discriminant analysis (LDA) classifiers. The photometric stereo method was used to acquire 2.5D depth maps.