Synergistic optimization of camera-aware disentanglement and consistency learning for unsupervised person re-identification

Qing Tian

doi:10.1016/j.neunet.2026.109448

Synergistic optimization of camera-aware disentanglement and consistency learning for unsupervised person re-identification

2026-08-06

0

PRE

AI

Q

Qing Tian *

L

Long Chen

B

Binghui Zhang

J

Jiashuo Shen

K

Keyang Cheng

Y

Yuhui Zheng

Z

Zhen Lei

DOI：10.1016/j.neunet.2026.109448

Original

Original request for help

Save

Abstract

En 中文

• SCC framework breaks the feature-label vicious cycle via co-optimization. • Camera-aware disentanglement mitigates bias via dynamic prototype learning. • Confidence-aware neighbor-center consistency refines noisy pseudo-labels. • Quantitative analyses verify camera invariance and pseudo-label reliability.

Keywords:

Unsupervised person re-identification

Synergistic optimization

Camera-aware representation learning

Pseudo-label refinement

Contrastive learning

Synergistic optimization of camera-aware disentanglement and consistency learning for unsupervised person re-identification

Abstract

Journal

Organization

Cited Papers

Citing Papers