Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Parallel Augmentation and Dual Enhancement for Occluded Person Re-identification

Zi Wang, Huaibo Huang, Aihua Zheng, Chenglong Li, Ran He

TL;DR

This work designs a parallel augmentation mechanism (PAM) to generate more suitable occluded data to mitigate the negative effects of unbalanced data and proposes the global and local dual enhancement strategy (DES) to promote the context information and details.

Abstract

Occluded person re-identification (Re-ID), the task of searching for the same person's images in occluded environments, has attracted lots of attention in the past decades. Recent approaches concentrate on improving performance on occluded data by data/feature augmentation or using extra models to predict occlusions. However, they ignore the imbalance problem in this task and can not fully utilize the information from the training data. To alleviate these two issues, we propose a simple yet effective method with Parallel Augmentation and Dual Enhancement (PADE), which is robust on both occluded and non-occluded data and does not require any auxiliary clues. First, we design a parallel augmentation mechanism (PAM) to generate more suitable occluded data to mitigate the negative effects of unbalanced data. Second, we propose the global and local dual enhancement strategy (DES) to promote the context information and details. Experimental results on three widely used occluded datasets and two non-occluded datasets validate the effectiveness of our method. The code is available at https://github.com/littleprince1121/PADE_Parallel_Augmentation_and_Dual_Enhancement_for_Occluded_Person_ReID

Parallel Augmentation and Dual Enhancement for Occluded Person Re-identification

TL;DR

This work designs a parallel augmentation mechanism (PAM) to generate more suitable occluded data to mitigate the negative effects of unbalanced data and proposes the global and local dual enhancement strategy (DES) to promote the context information and details.

Abstract

Occluded person re-identification (Re-ID), the task of searching for the same person's images in occluded environments, has attracted lots of attention in the past decades. Recent approaches concentrate on improving performance on occluded data by data/feature augmentation or using extra models to predict occlusions. However, they ignore the imbalance problem in this task and can not fully utilize the information from the training data. To alleviate these two issues, we propose a simple yet effective method with Parallel Augmentation and Dual Enhancement (PADE), which is robust on both occluded and non-occluded data and does not require any auxiliary clues. First, we design a parallel augmentation mechanism (PAM) to generate more suitable occluded data to mitigate the negative effects of unbalanced data. Second, we propose the global and local dual enhancement strategy (DES) to promote the context information and details. Experimental results on three widely used occluded datasets and two non-occluded datasets validate the effectiveness of our method. The code is available at https://github.com/littleprince1121/PADE_Parallel_Augmentation_and_Dual_Enhancement_for_Occluded_Person_ReID
Paper Structure (12 sections, 6 equations, 2 figures, 5 tables)

This paper contains 12 sections, 6 equations, 2 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Proposed Methods
  3. Parallel Augmentation Mechanism
  4. Dual Enhancement Strategy
  5. Loss Function
  6. Implementation Details
  7. Experiment
  8. Results on Occluded Re-ID
  9. Results on non-occluded Re-ID
  10. Ablation Study
  11. Robustness Evaluation on Occluded Data
  12. Conclusion

Figures (2)

  • Figure 1: (a) & (b): Imbalance problem. (c) & (d): Global and local information have their advantages, respectively.
  • Figure 2: Overall structure of PADE. First, we implement erase and crop operations on original inputs to form the image triplet (Original, Erased, and Cropped images). The image triplet will be sent to the ViT-based backbone to extract global features. Then the global and local features from the non-occluded image branch will be interactively enhanced by each other.