Face Recognition of the Rhinopithecus Roxellana Qinlingensis Based on Improved HOG and Sparse Representation

With the researches on face recognition of Rhinopithecusroxellanaqinlingensis, this thesis comes up with some methods that refining traditional HOG and Sparse Representation in order to improve the efficiency in recognizing golden monkeys. As we know, improved HOG is an optimal way to show partial information of an image. Besides, it can also plays an crucial role in staying stability in both optical and geometric distortion, which means the changes in expressions, postures and angles of golden monkeys can also be ignored. By using these characteristics as a alternation of original images to be a part of Sparse dictionary, and make a facial recognition on golden monkey with Sparse Representation, which can be a ideal method to erase many unnecessary messages and improve the accuracy on facial recognition of golden monkeys. Compared with mainstream method in recognition, this method is more reliable and effective and has a higher efficiency in recognition.