WCSE

A Specific Face Positioning Segmentation Research Based on Complex Background

Abstract— Facial gesture, light, expression, and other factors affect the formation of a complex background, under this wireless visual sensor network identification encountered unprecedented challenges. Traditional visual identification is required at node distribution is relatively complex environment, the uncontrollable factors transforming into a controllable, stable characteristic factors during the identification, thus leading to long recognition time, low efficiency. In this paper, the use of face recognition based on LBP wireless visual sensor network identity adaptive recognition method to get strong classifier cascade of underlying feature extraction, through the final Harr face cascade classifier human face testing, and the face region feature vector PCA dimension reduction, access to low-dimensional feature vector, the number of drop maintain, establish the simulation experiments, the feature matching facial feature library to complete the identification. Simulation results show that the proposed algorithm has high accuracy, effectively improve the recognition efficiency.

Index Terms— wireless sensor, dimensionality reduction, feature matching, adaptive recognition.

Wang Jiang-Tao
College of Engineering Technology, Yang-En University, CHINA
Chen Duan-Sheng, Wang Jing
College of Computer Science and Technolgy, Huaqiao University, CHINA

[Download]

Cite: Wang Jiang-Tao, Chen Duan-Sheng, Wang Jing, "A Specific Face Positioning Segmentation Research Based on Complex Background," Proceedings of 2016 6th International Workshop on Computer Science and Engineering, pp. 743-748, Tokyo, 17-19 June, 2016.