Face Identification Based On Local Binary Patterns: A Survey

The goal of this research project was to come up with a combined face recognition algorithm which out-performs existing algorithms on accuracy. The identification of individuals using face recognition techniques is a challenging task. This is due to the variations resulting from facial expressions, illuminations, makeup, rotations, and gestures. Facial images also contain a great deal of the redundant information, which negatively affects the performance of the recognition system. This paper proposes a novel approach for recognizing facial images from facial features using feature descriptors, namely local binary patterns (LBP) and local directional patterns (LDP). This research work consists of three parts, namely face representation, feature extraction and classification. The useful and unique features of the facial images have been extracted in the feature extraction phase. In the classification, the face image has been compared with the images from the database. The face area has been divided into small regions from which local binary and directional patterns (LBP/LDP) histograms have been extracted and concatenated into a single feature vector (histogram). Experiments performed on the Cohn-Kanade facial expression database have shown a good recognition rate 99% indicating superiority of the proposed method compared to other methods. The proposed methods includes a combination of local binary pattern(LBP) and local directional patterns(LDP+LBP+Voting Classifier) as the feature extractor and voting classifier classification algorithm which is an aggregate classifier composed of k-Nearest Neighbor, decision trees and support vector machines. The results reveal improved accuracy results as compared to other local binary pattern variants in both scenarios where small datasets or huge datasets are used.