EUVIMEDEUVIMED entdecken
Crossref

Cross-Modality Self-Distillation for Visible-Infrared Person Re-Identification

Zihao Hu, Yongguo Ling, Zhiming Luo, Wenhao Shao, Shaozi Li, Zhun Zhong

IEEE Transactions on Geoscience and Remote Sensing · 2026 · 64 · 5624315-5624315

DOI: 10.1109/tgrs.2026.3696321 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Geoscience and Remote Sensing, 2026. Verfasst von Zihao Hu und 5 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 85 Literaturangaben. Gefördert von: Guangxi Natural Science Foundation of China, National Natural Science Foundation of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Visible-infrared person re-identification (VI-ReID) faces large challenges due to the cross-modality gap and intraclass variations. Existing methods mainly focus on mitigating the cross-modality discrepancy by optimizing the embedding space of high-level features. However, two critical issues are often overlooked: low-level features suffer from larger modality gaps and redundant channels exhibit limited identity discrimination. These limitations hinder the overall discriminative performance of the model. To address the above issues, we propose a novel cross-modality self-distillation (CMSD) network, which consists of two key components: cross-layer semantic distillation (CSD) and intralayer alignment distillation (IAD). These two modules are mutually beneficial to jointly explore cross-layer modality discrepancy and intralayer channel discrepancy. Specifically, CSD distills modality-shared semantics from high-level features to guide low-level features in learning more reliable modality-invariant representations. Furthermore, IAD decomposes feature maps into salient and redundant channels while transferring identity-related information from salient to redundant channels, thereby enhancing overall discrimination of modality-invariant features. Our extensive experiments on the SYSU-MM01, RegDB, and LLCM datasets demonstrate that the components of our proposed CMSD mutually enhance each other, leading to superior performance compared to state-of-the-art (SOTA) methods.

Förderung

  • Guangxi Natural Science Foundation of China
  • National Natural Science Foundation of China

Zitieren

APA

Hu, Z., Ling, Y., Luo, Z., Shao, W., Li, S., & Zhong, Z. (2026). Cross-Modality Self-Distillation for Visible-Infrared Person Re-Identification. IEEE Transactions on Geoscience and Remote Sensing, 64, 5624315-5624315. https://doi.org/10.1109/tgrs.2026.3696321

Vancouver

Hu Z, Ling Y, Luo Z, Shao W, Li S, Zhong Z. Cross-Modality Self-Distillation for Visible-Infrared Person Re-Identification. IEEE Transactions on Geoscience and Remote Sensing. 2026;64:5624315-5624315. doi:10.1109/tgrs.2026.3696321

BibTeX
@article{Hu2026,
  author = {Hu, Zihao and Ling, Yongguo and Luo, Zhiming and Shao, Wenhao and Li, Shaozi and Zhong, Zhun},
  title = {Cross-Modality Self-Distillation for Visible-Infrared Person Re-Identification},
  journal = {IEEE Transactions on Geoscience and Remote Sensing},
  year = {2026},
  volume = {64},
  pages = {5624315--5624315},
  doi = {10.1109/tgrs.2026.3696321}
}
Kostenlos starten →