Study On Gender And Expression Recognition Algorithms Based On Facial Features

Gender and facial expression recognition have received substantial attention among the researchers in the domain of pattern recognition,artificial intelligence and applied mathematics,according to the application in many fields such as human-computer interaction,safe driving and surveillance.But algorithm of gender and facial expression recognition can be easily influenced by the pose and illumination,and lack of generalization.In this paper,a new method based on facial features for gender and facial expression is proposed.First,we use block Local Binary Pattern(LBP)as feature,which can highlight the differences between different parts of the face and,it is robust to illumination and rotation with good distinguishing ability.Then,the gender and facial expression recognition algorithm based on Gradient Boosting Decision Tree(GBDT)algorithm is proposed.GBDT algorithm is comprised of decision tree and gradient boosting frame,and it can be applied to solve the problems of regression and classification.In this paper,we use it in gender and facial expression recognition of facial images.it is the first time the proposed algorithm has been used in gender and facial expression recognition of facial images.GBDT algorithm can be directly applied in multiple classifications,and it has the advantage of handling large scale data and has better generalization ability compared with other classification algorithms.Finally,experiments in self-built gender database of the application mathematics of NEU and Cohn-Kanade++ database show that the recognition rate are 94.4%and 93.1%.At the same time,in order to verify the generalization ability of the algorithm,we conduct experiments on the FERET and JAFFE respectively,the recognition rate are 92%and 85%,and the expected results were achieved.