Face Recognition with Local Binary Patterns, Spatial Pyramid Histograms and Nearest Neighbor Classification

Face recognition algorithms generally assume that face images are well aligned and have a similar pose yet in many different practical applications it is impossible to meet these certain conditions. Thus extending face recognition to unconstrained face images has become an active area for research. At this end, histograms of Local Binary Patterns (LBP) have proven to be highly discriminative descriptors for face recognition. Most LBP-based algorithms use a rigid descriptor matching strategy thats not robust against pose variation and misalignment. Here two algorithms are proposed for face recognition which are designed to deal with pose variations and misalignment. It also incorporate an illumination normalization step that increases robustness against lighting variations. The proposed algorithms use descriptors based on histograms of LBP and perform descriptor matching with spatial pyramid matching (SPM) and Naive Bayes Nearest Neighbor (NBNN) respectively. The main contribution is the inclusion of flexible spatial matching schemes, it uses an image-to-class relation to provide an improved robustness with respect to intra-class variations. The comparison is compulsory between the accuracy of the proposed algorithms against Ahonens original LBP-based face recognition system and two baseline holistic classifiers on four standard datasets. Results indicate that the algorithm based on NBNN outperforms the other solutions and does so more markedly in presence of pose variations.