Improving Person Re-identification by Attribute and Identity Learning

By Yutian Lin, Liang Zheng, Zhedong Zheng, Yu Wu, Yi Yang. Pattern Recognition, 2017

881 citations recorded by the source corpus.

Abstract Person re-identification (re-ID) and attribute recognition share a common target at learning pedestrian descriptions. Their difference consists in the granularity. Most existing re-ID methods only take identity labels of pedestrians into consideration. However, we find the attributes, containing detailed local descriptions, are beneficial in allowing the re-ID model to learn more discriminative feature representations. In this paper, based on the complementarity of attribute labels and ID labels, we propose an attribute-person recognition (APR) network, a multi-task network which learns a re-ID embedding and at the same time predicts pedestrian attributes. We manually annotate attribute labels for two large-scale re-ID datasets, and systematically investigate how person re-ID and attribute recognition benefit from each other. In addition, we re-weight the attribute predictions considering the dependencies and correlations among the attributes. The experimental results on two large-scale re-ID benchmarks demonstrate that by learning a more discriminative representation, APR achieves competitive re-ID performance compared with the state-of-the-art methods. We use APR to speed up the retrieval process by ten times with a minor accuracy drop of 2.92% on Market-1501. Besides, we also apply APR on the attribute recognition task and demonstrate improvement over the baselines.

Subfields: Biomedical Engineering, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/7f23a4bb0c777dd72cca7665a5f370ac7980217e

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. Zhedong Zheng
4. Qi Tian
5. Mang Ye
6. Wanli Ouyang
7. Shaogang Gong
8. Tao A. Xiang

In Computer Vision and Pattern Recognition

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

Citations in the corpus

• Multi-level Factorisation Net for Person Re-identification (2018, background): “These results thus show that our low-dimensional MLFN-Ŝ alone can be more effective in attribution prediction than APR. Remind that MLFN is trained without annotated attributes while APR network is designed for supervised attribute learning.”
• AlignedReID: Surpassing Human-Level Performance in Person Re-Identification (2017, background): “With the renaissance of deep learning, the convolutional neural network (CNN) has dominated this field [24, 32, 6, 54, 16, 24], by learning features in an end-to-end fashion through various metric learning losses such as contrastive loss [32], triplet loss [18], improved triplet loss [6], quadruplet”
• Transferable Joint Attribute-Identity Deep Learning for Unsupervised Person Re-identification (2018, methodology): “Attribute Annotation In our evaluations, we use either Market-1501 [56] or DukeMTMC-ReID) dataset [30] as the source domain, since they provide both identity and attribute labels (Fig.”
• In Defense of the Triplet Loss for Person Re-Identification (2017, background): “Several authors suggest learning attributes and ReID jointly [17, 33, 24, 30], some of which integrate this into their embedding dimensions.”
• AANet: Attribute Attention Network for Person Re-Identifications (2019, methodology): “[15] also annotated this dataset, but with 27 person attributes.”
• Pose-Normalized Image Generation for Person Re-identification (2017, methodology): “The predictor is trained on the attribute labels of Market-1501 dataset training split, annotated by [34].”
• Person Re-identification by Deep Learning Multi-scale Representations (2017, background): “2 As this dataset was newly constructed for person re-id from the multitarget multi-camera tracking benchmark DukeMTMC [46], there are only a small number of results reported in a few unpublished arXiv papers [78, 31, 52].”
• Random Erasing Data Augmentation (2017, neutral): “[8] 86 [32] 25.13 12.17 LOMO + XQDA [16] 30.75 17.04 IDE [33] 65.22 44.99 GAN [35] 67.68 47.13 OIM [25] 68.1 47.4 APR [17] 70.69 51.88 PAN [34] 71.59 51.51 PAN + re-rank [34] 75.94 66.74 SVDNet [21] 76 Table 10.”
• Part-Aligned Bilinear Representations for Person Re-identification (2018, methodology): “We follow the setting in [70] to conduct the experiments.”
• Exploit the Unknown Gradually: One-Shot Video-Based Person Re-identification by Stepwise Learning (2018, methodology): “Following recent works [7, 17, 38], we take the training process as an identity classification task.”
• Bi-Directional Center-Constrained Top-Ranking for Visible Thermal Person Re-Identification (2020, background): “In addition, semantic attributes information is also aggregated with visible images to improve the person re-identification performance [39], [40].”
• Pose-Guided Feature Alignment for Occluded Person Re-Identification (2019, neutral): “We [42] 81.0 63.4 - - Pedestrian[47] 82.0 63.0 - - DSR [11] 83.5 64.2 - - MultiLoss [35] 83.9 64.4 - - TripletLoss [13] 84.9 69.1 - - Adver Occluded [14] 86.5 78.3 79.1 62.1 APR [22] 87.0 66.9 73.9 55.6 MultiScale [4] 88.9 [32].”
• Enhancing Person Re-identification in a Self-Trained Subspace (2017, background): “During the past decades, many person re-ID algorithms [2], [3], [11], [13], [14], [21]–[26] have been proposed.”