Person Re-identification with Correspondence Structure Learning
This addresses the problem of improving accuracy in person re-identification for surveillance systems, but it appears incremental as it builds on existing methods for handling misalignments.
The paper tackles spatial misalignments in person re-identification caused by camera-view changes or human-pose variations by learning a correspondence structure for patch-wise matching and integrating a global matching constraint, achieving effective results on various datasets.
This paper addresses the problem of handling spatial misalignments due to camera-view changes or human-pose variations in person re-identification. We first introduce a boosting-based approach to learn a correspondence structure which indicates the patch-wise matching probabilities between images from a target camera pair. The learned correspondence structure can not only capture the spatial correspondence pattern between cameras but also handle the viewpoint or human-pose variation in individual images. We further introduce a global-based matching process. It integrates a global matching constraint over the learned correspondence structure to exclude cross-view misalignments during the image patch matching process, hence achieving a more reliable matching score between images. Experimental results on various datasets demonstrate the effectiveness of our approach.