Pyramidal person re-identification via multi-loss dynamic training

Zheng，Feng1; Deng，Cheng2; Sun，Xing3; Jiang，Xinyang3; Guo，Xiaowei3; Yu，Zongqiao3; Huang，Feiyue3; Ji，Rongrong4,5

10.1109/CVPR.2019.00871

2019-06-01

1063-6919

978-1-7281-3294-5

Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition

2019-June

8506-8514

15-20 June 2019

Long Beach, CA, United states

345 E 47TH ST, NEW YORK, NY 10017 USA

IEEE Computer Society

Most existing Re-IDentification (Re-ID) methods are highly dependent on precise bounding boxes that enable images to be aligned with each other. However, due to the challenging practical scenarios, current detection models often produce inaccurate bounding boxes, which inevitably degenerate the performance of existing Re-ID algorithms. In this paper, we propose a novel coarse-to-fine pyramid model to relax the need of bounding boxes, which not only incorporates local and global information, but also integrates the gradual cues between them. The pyramid model is able to match at different scales and then search for the correct image of the same identity, even when the image pairs are not aligned. In addition, in order to learn discriminative identity representation, we explore a dynamic training scheme to seamlessly unify two losses and extract appropriate shared information between them. Experimental results clearly demonstrate that the proposed method achieves the state-of-the-art results on three datasets. Especially, our approach exceeds the current best method by 9.5% on the most challenging CUHK03 dataset.

Categorization Deep Learning Motion and Tracking Recognition: Detection Representati Retrieval Vision Applications and Systems

第一

英语

EI ; CPCI-S

[2017KSYS008]

Computer Science

Computer Science, Artificial Intelligence ; Computer Science, Theory & Methods

WOS:000542649302013

20200508114551

Motion analysis ; Computer vision

Ergonomics and Human Factors Engineering:461.4 ; Data Processing and Image Processing:723.2 ; Computer Applications:723.5 ; Vision:741.2

2-s2.0-85076555957

Scopus

1.Department of Computer Science and Engineering,Southern University of Science and Technology,Shenzhen,518055,China
2.School of Electronic Enigineering,Xidian University,Xi'an,710071,China
3.YouTu Lab Tencent,Shanghai,China
4.School of Information Science and Engineering,Xiamen University,Xiamen,China
5.Peng Cheng Laboratory,Shenzhen,China

Zheng，Feng,Deng，Cheng,Sun，Xing,et al. Pyramidal person re-identification via multi-loss dynamic training[C]. 345 E 47TH ST, NEW YORK, NY 10017 USA:IEEE Computer Society,2019:8506-8514.