Batch DropBlock Network for Person Re-identification and Beyond
This addresses pose and occlusion issues in person re-identification, offering a method applicable to general metric learning tasks, but it is incremental as it builds on existing CNN architectures.
The paper tackles the problem of pose changes and occlusions in person re-identification by proposing a Batch DropBlock Network that uses a two-branch approach to reinforce attentive feature learning, achieving state-of-the-art results such as 76.4% Rank-1 accuracy on CUHK03-Detect and 83.0% Recall-1 on Stanford Online Products.
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%).