Visibility-Guided and Occlusion-Simulated Learning for Robust Person Re-Identification
Junjie Cao, Rong Rong, Xing Xie · 2026
Read Visibility-Guided and Occlusion-Simulated Learning for Robust Person Re-Identification by Junjie Cao, Rong Rong, Xing Xie (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13210425.txt.
Description
Junjie Cao, Rong Rong, and Xing Xie built a person re-identification system in 2026 that copes with blocked views. Matching people across cameras fails when bags, cars, or crowds hide body parts. Their framework judges which parts are visible and practises on fake blockages during training.
One module scores the reliability of each body part and weights features toward visible parts. A second module pastes simulated occlusion boxes over training images, teaching robustness under several supervision branches. On five benchmarks, occluded and normal alike, the system reached competitive accuracy. Code is public. The authors aim at surveillance and retail settings where partial views are the norm.
Catalog record
| CREATOR | Junjie Cao, Rong Rong, Xing Xie |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 42 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |