Learning Deep Context-Aware Features over Body and Latent Parts for Person Re-identification

By Dangwei Li, Xiaotang Chen, Zhang Zhang, Kaiqi Huang. Computer Vision and Pattern Recognition, 2017

704 citations recorded by the source corpus.

Person Re-identification (ReID) is to identify the same person across different cameras. It is a challenging task due to the large variations in person pose, occlusion, background clutter, etc. How to extract powerful features is a fundamental problem in ReID and is still an open problem today. In this paper, we design a Multi-Scale Context-Aware Network (MSCAN) to learn powerful features over full body and body parts, which can well capture the local context knowledge by stacking multi-scale convolutions in each layer. Moreover, instead of using predefined rigid parts, we propose to learn and localize deformable pedestrian parts using Spatial Transformer Networks (STN) with novel spatial constraints. The learned body parts can release some difficulties, e.g. pose variations and background clutters, in part-based representation. Finally, we integrate the representation learning processes of full body and body parts into a unified framework for person ReID through multi-class person identification tasks. Extensive evaluations on current challenging large-scale person ReID datasets, including the image-based Market1501, CUHK03 and sequence-based MARS datasets, show that the proposed method achieves the state-of-the-art results.

Subfields: Biomedical Engineering, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/605886b07ec27492e8d5bc48f79ddc39c02ec190

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Biomedical Engineering

1. Yi Yang
2. Liang Zheng
3. Qi Tian
4. Wanli Ouyang
5. Tao A. Xiang
6. Shiliang Zhang
7. Zhedong Zheng
8. Shaogang Gong

In Computer Vision and Pattern Recognition

1. Yi Yang
2. Luc Van Gool
3. Dacheng Tao
4. Xiaogang Wang
5. Ming-Hsuan Yang
6. Tao A. Xiang
7. Ling Shao
8. Liang Zheng

Citations in the corpus

• TransReID: Transformer-based Object Re-Identification (2021, background): “Among them, local part features [37, 44, 20, 48, 28] and side information (such as cameras and viewpoints) [7, 61, 35, 30], have been proven to be essential and effective to enhance the feature robustness.”
• Pose-Normalized Image Generation for Person Re-identification (2017, background): “A number of models [39, 62, 58, 57, 23, 50, 45] are proposed to address this problem.”
• Diversity Regularized Spatiotemporal Attention for Video-Based Person Re-identification (2018, background): “In recent years, researchers have proposed numerous deep learning based methods [1, 24, 8, 20, 44] to jointly handle both aspects.”
• Human Semantic Parsing for Person Re-identification (2018, methodology): “[21] learn the body parts roughly as head-shoulder, upper-body and lower-body using a spatial transformer network [19].”
• Improving Person Re-identification by Attribute and Identity Learning (2017, neutral): “87 Cont-aware [50] CVPR2017 * 80.”
• A Bi-Directional Message Passing Model for Salient Object Detection (2018, methodology): “[13] embed the multi-scale contextual information by stacking sequential blocks containing several dilated convolutions and use the features from the last block for prediction.”
• Person Transfer GAN to Bridge Domain Gap for Person Re-identification (2017, background): “Similarly, Li et al. [18] adopt Spatial Transform Networks (STN) [13] to detect latent part regions and then learn descriptors on those regions.”
• A Pose-Sensitive Embedding for Person Re-identification with Expanded Cross Neighborhood Re-ranking (2017, neutral): “7 - Latent Parts [16] CVPR17 57.”
• Deep Cosine Metric Learning for Person Re-identification (2018, neutral): “60 MSCAN [12] 71.”
• Interaction-And-Aggregation Network for Person Re-Identification (2019, neutral): “Person reID methods focus on two key points: learning a powerful feature representation for images [8, 43, 14, 44, 24, 26, 28, 42, 1] and designing an effective distance metric [32, 29, 56, 2, 57].”
• Exploit the Unknown Gradually: One-Shot Video-Based Person Re-identification by Stepwise Learning (2018, methodology): “One simple solution is using image-based re-ID methods, and obtaining video representations by pooling the frame features [10, 16, 20].”
• Joint Discriminative and Generative Learning for Person Re-Identification (2019, neutral): “Market-1501 DukeMTMC-reID Rank@1 mAP Rank@1 mAP Verif-Identif [55] 79.5 59.9 68.9 49.3 DCF [22] 80.3 57.5 - - SSM [2] 82.2 68.8 - - SVDNet [38] 82.3 62.1 76.7 56.8 PAN [57] 82.8 63.4 71.6 51.5 GLAD [47] 89.9 73.9 - - HA-CNN [24] 91 two training phases to sequentially train generative models and…”
• Mancs: A Multi-task Attentional Network with Curriculum Sampling for Person Re-Identification (2018, result): “Attention is very useful in person re-ID, which has been proved in the previous studies [15, 16, 20].”
• Harmonious Attention Network for Person Re-identification (2018, background): “A small number of attention deep learning models for re-id have been recently developed for reducing the negative effect from poor detection and human pose change [17, 42, 27].”
• Multi-level Factorisation Net for Person Re-identification (2018, neutral): “6 - Context [22] 80.”
• Visual Attention Methods in Deep Learning: An In-Depth Survey (2022, background): “to tackle the limitation of the previous attention modules in person Re-Identification by learning attention selection and feature representation jointly and hence solving the issue of misalignment calibration caused by constrained attention mechanisms [33], [34], [35], [36].”
• Reinforced Temporal Attention and Split-Rate Transfer for Depth-Based Person Re-identification (2017, background): “Other approaches handle the problem of pose variability by explicitly accounting for spatial constraints of the human body parts [12,97,39,96] or by predicting the pose from video [16,72].”