Batch DropBlock Network for Person Re-identification and Beyond
Authors:
Preprint
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%).Keywords
All Related Versions
This publication has 62 references indexed in Scilit:
- Part-Aligned Bilinear Representations for Person Re-identification2018
- Deep Metric Learning with Hierarchical Triplet Loss2018
- Attention-Based Ensemble for Deep Metric Learning2018
- Deep Randomized Ensembles for Metric Learning2018
- Beyond Part Models: Person Retrieval with Refined Part Pooling (and A Strong Convolutional Baseline)2018
- Performance Measures and a Data Set for Multi-target, Multi-camera Tracking2016
- Gated Siamese Convolutional Neural Network Architecture for Human Re-identification2016
- A Siamese Long Short-Term Memory Architecture for Human Re-identification2016
- Consistent Re-identification in a Camera Network2014
- Salient Color Names for Person Re-identification2014