Person re-identification by Weber Local Descriptors and color features

Person re-identification plays a key role in video monitoring. Aiming for current person re-identifications for numerical complexity and extraction difficulty, we propose a simple and fast multi-feature. On the basis of analysis of difference excitation and orientation of Weber Local Descriptors, showing graphic texture features by difference excitation of circular field, showing graphic edge orientation by orientation of Local Binary Pattern coding, then showing graphic color features by HSV color space histogram, last integrate all features. Experimental results show that multi-features work fast and has a strong robustness on changed posture, angles, light and shielded body part within ETHZ, VIPeR person re-identification datasets.