Batch DropBlock Network for Person Re-identification and Beyond

TL;DR

The paper introduces Batch DropBlock Network, a two-branch CNN architecture that enhances person re-identification by combining global features with attentive local features through a novel feature dropping technique, achieving state-of-the-art results.

Contribution

It proposes Batch DropBlock, a simple yet effective feature dropping method integrated into a two-branch network for improved local feature learning in re-identification.

Findings

Achieves 76.4% Rank-1 accuracy on CUHK03-Detect

Achieves 83.0% Recall-1 on Stanford Online Products

Outperforms existing methods by over 6% in key metrics

Abstract

Since the person re-identification task often suffers from the problem of pose changes and occlusions, some attentive local features are often suppressed when training CNNs. In this paper, we propose the Batch DropBlock (BDB) Network which is a two branch network composed of a conventional ResNet-50 as the global branch and a feature dropping branch. The global branch encodes the global salient representations. Meanwhile, the feature dropping branch consists of an attentive feature learning module called Batch DropBlock, which randomly drops the same region of all input feature maps in a batch to reinforce the attentive feature learning of local regions. The network then concatenates features from both branches and provides a more comprehensive and spatially distributed feature representation. Albeit simple, our method achieves state-of-the-art on person re-identification and it is also applicable to general metric learning tasks. For instance, we achieve 76.4% Rank-1 accuracy on the CUHK03-Detect dataset and 83.0% Recall-1 score on the Stanford Online Products dataset, outperforming the existing works by a large margin (more than 6%).