Race recognition with Local Binary Pattern

Automatic recognition of races is major component of human-machine interactions. Recognition of races can reduce number of images that captured security camera to find any criminal whose race is known. It is useful especially for airports, border gates and other social domains. In this paper, three major races class which are Caucasion, Mongoloid and Negroid are observed. Local binary patterns (LBP) are used for feature extraction process, and weighted chi square statistics are used for recognition process. To classify races, models that are created by using morphing are used. The test results show that the system performance is approximately 99%.