Partial Person Re-identification Method Based on Visible Perceptual Texture Semantic Alignment

Abstract

A partial pedestrian re-identification method (TSA) based on visible-aware texture semantic alignment. This method can efficiently solve the two problems of pedestrian occlusion and change of posture or viewing angle at the same time. The specific steps of the method are as follows: (1) Design a local area alignment network based on human posture, focusing on solving the problem of pedestrians being occluded; (2) Designing a texture alignment network based on semantic visibility, focusing on solving the problem of pedestrian posture changes or perspective changes; (3) In order to make the model have better generalization ability, the two networks are jointly learned, so that the model can better deal with pedestrian occlusion and changes in posture or observation perspective. The present invention performs efficient partial pedestrian re-identification based on visible perception texture semantic alignment and partial area alignment based on human posture, which can effectively solve the problems of occlusion and posture change in pedestrian re-identification, and the method has a fast convergence speed and can be implemented in pedestrian occlusion. Efficient re-identification.

Description

technical field [0001] The invention belongs to the technical field of computer vision and pattern recognition, and relates to a partial pedestrian re-identification method (TSA) based on visible perception texture semantic alignment, which can align on texture and local area at the same time, and can solve the problem of occlusion and occlusion in pedestrian re-identification. The problem of attitude transformation. Background technique [0002] In recent years, with the continuous improvement of infrastructure construction, in order to ensure public security and safety, tens of millions of non-overlapping (disjoint) cameras have been placed in every corner of many cities. In some special cases, when the target person disappears from one camera, we want to quickly re-identify the target person from other cameras. This is exactly the research hotspot in the field of computer vision and machine learning - person re-identification (ReID) task. Because of its importance in pu...

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to solve the problem of inconsistency of input image scales in the existing classical algorithms SWM[1] and AWC[1], the problem of non-shared area feature interference in the algorithms of DSR[2] and DCR[3], and the problems not considered in these methods To the problem of human body pose transformation in real scenes, a partial pedestrian re-identification method (TSA) based on visible perceptual texture semantic alignment is provided

Images