ReMamba: a hybrid CNN-Mamba aggregation network for visible-infrared person re-identification
Haokun Geng, Jiaren Peng, Wenzhong Yang, Danny Chen, Hongzhen Lv, Guanghan Li, Yi Shao · 2024
Read ReMamba: a hybrid CNN-Mamba aggregation network for visible-infrared person re-identification by Haokun Geng, Jiaren Peng, Wenzhong Yang, Danny Chen, Hongzhen Lv, Guanghan Li, Yi Shao (2024) 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-11599763.txt.
Description
Haokun Geng and six co-authors build a system that can match photos of the same person taken by normal cameras and by infrared cameras. The task, called visible-infrared person re-identification, is hard because the two image types look quite different, and one person can vary a lot across views. Earlier systems based on plain convolutional networks miss global context, the authors note.
Their network, ReMamba, starts with a convolutional backbone that pulls features at several levels. A state space model then blends local detail into the global picture. An adaptive module filters and merges the two kinds of features, guided by an extra loss term. A second loss term pushes features from the two camera types closer together. On three public datasets, ReMamba beat the best prior methods, the authors report.
Catalog record
| CREATOR | Haokun Geng, Jiaren Peng, Wenzhong Yang, Danny Chen, Hongzhen Lv, Guanghan Li, Yi Shao |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 89 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |