Dynamic Feature Pruning and Consolidation for Occluded Person Re-Identification

TL;DR

This paper introduces a feature pruning and consolidation framework for occluded person re-identification that avoids explicit human structure parsing and effectively handles heavy occlusions, outperforming existing methods.

Contribution

The proposed FPC framework uses a sparse encoder and feature consolidation to improve occluded person ReID without relying on human structure cues.

Findings

Outperforms state-of-the-art by at least 8.6% mAP on Occluded-Duke.

Effectively handles heavy occlusion and partial re-identification.

Demonstrates robustness across multiple occluded and holistic datasets.

Abstract

Occluded person re-identification (ReID) is a challenging problem due to contamination from occluders. Existing approaches address the issue with prior knowledge cues, such as human body key points and semantic segmentations, which easily fail in the presence of heavy occlusion and other humans as occluders. In this paper, we propose a feature pruning and consolidation (FPC) framework to circumvent explicit human structure parsing. The framework mainly consists of a sparse encoder, a multi-view feature mathcing module, and a feature consolidation decoder. Specifically, the sparse encoder drops less important image tokens, mostly related to background noise and occluders, solely based on correlation within the class token attention. Subsequently, the matching stage relies on the preserved tokens produced by the sparse encoder to identify k-nearest neighbors in the gallery by measuring the image and patch-level combined similarity. Finally, we use the feature consolidation module to compensate pruned features using identified neighbors for recovering essential information while disregarding disturbance from noise and occlusion. Experimental results demonstrate the effectiveness of our proposed framework on occluded, partial, and holistic Re-ID datasets. In particular, our method outperforms state-of-the-art results by at least 8.6\% mAP and 6.0\% Rank-1 accuracy on the challenging Occluded-Duke dataset.