Camera-Agnostic Person Re-Identification via Adversarial Disentangling Learning

Ni，Hao1; Song，Jingkuan1; Zhu，Xiaosu1; Zheng，Feng2; Gao，Lianli1

10.1145/3474085.3475361

2021-10-17

MM 2021 - Proceedings of the 29th ACM International Conference on Multimedia

2002-2010

Despite the success of single-domain person re-identification (ReID), current supervised models degrade dramatically when deployed to unseen domains, mainly due to the discrepancy across cameras. To tackle this issue, we propose an Adversarial Disentangling Learning (ADL) framework to decouple camera-related and ID-related features, which can be readily used for camera-agnostic person ReID. ADL adopts a discriminative way instead of the mainstream generative styles in disentangling methods, eg., GAN or VAE based, because for person ReID task only the information to discriminate IDs is needed, and more information to generate images are redundant and may be noisy. Specifically, our model involves a feature separation module that encodes images into two separate feature spaces and a disentangled feature learning module that performs adversarial training to minimize mutual information. We design an effective solution to approximate and minimize mutual information by transforming it into a discrimination problem. The two modules are co-designed to obtain strong generalization ability by only using source dataset. Extensive experiments on three public benchmarks show that our method outperforms the state-of-the-art generalizable person ReID model by a large margin. Our code is publicly available at https://github.com/luckyaci/ADL_ReID.

disentanglement generalizable person re-identification neural networks

其他

英语

EI

20214711200008

Computer vision

Computer Applications:723.5 ; Vision:741.2 ; Photographic Equipment:742.2

2-s2.0-85119347886

Scopus

1.University of Electronic Science and Technology of China,Chengdu,China
2.Southern University of Science and Technology,Shenzhen,China

Ni，Hao,Song，Jingkuan,Zhu，Xiaosu,et al. Camera-Agnostic Person Re-Identification via Adversarial Disentangling Learning[C],2021:2002-2010.