EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

TASA: Text-Anchored State–Space Alignment for Long-Tailed Image Classification

Long Li iD, Tinglei Jia iD, Huaizhi Yue iD, Huize Cheng iD, Yongfeng Bu iD, Zhaoyang Zhang iD

Journal of Imaging · 13.11.2025 · 11 (11) · 410

DOI: 10.3390/jimaging11110410 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Imaging, online erschienen am 13.11.2025. Verfasst von Long Li und 5 weiteren Autor:innen. Beteiligte Einrichtung: School of Information Engineering, Chang'an University, Xi'an 710064, China. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 58 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Long-tailed image classification remains challenging for vision–language models. Head classes dominate training while tail classes are underrepresented and noisy, and short prompts with weak text supervision further amplify head bias. This paper presents TASA, an end-to-end framework that stabilizes textual supervision and enhances cross-modal fusion. A Semantic Distribution Modulation (SDM) module constructs class-specific text prototypes by cosine-weighted fusion of multiple LLM-generated descriptions with a canonical template, providing stable and diverse semantic anchors without training text parameters. Dual-Space Cross-Modal Fusion (DCF) module incorporates selective-scan state–space blocks into both image and text branches, enabling bidirectional conditioning and efficient feature fusion through a lightweight multilayer perceptron. Together with a margin-aware alignment loss, TASA aligns images with class prototypes for classification without requiring paired image–text data or per-class prompt tuning. Experiments on CIFAR-10/100-LT, ImageNet-LT, and Places-LT demonstrate consistent improvements across many-, medium-, and few-shot groups. Ablation studies confirm that DCF yields the largest single-module gain, while SDM and DCF combined provide the most robust and balanced performance. These results highlight the effectiveness of integrating text-driven prototypes with state–space fusion for long-tailed classification.

Schlagwörter

Stichwörter der Autor:innen

cross-modal alignmentlong-tailed image classificationtext prototypesvision–language models

Autor:innen und Institutionen

  • School of Information Engineering, Chang'an University, Xi'an 710064, China.

ORCID-Profile: Long Li, Tinglei Jia, Huaizhi Yue, Huize Cheng, Yongfeng Bu, Zhaoyang Zhang

Zitieren

APA

Li, L., Jia, T., Yue, H., Cheng, H., Bu, Y., & Zhang, Z. (2025). TASA: Text-Anchored State–Space Alignment for Long-Tailed Image Classification. Journal of Imaging, 11(11), 410. https://doi.org/10.3390/jimaging11110410

Vancouver

Li L, Jia T, Yue H, Cheng H, Bu Y, Zhang Z. TASA: Text-Anchored State–Space Alignment for Long-Tailed Image Classification. Journal of Imaging. 2025;11(11):410. doi:10.3390/jimaging11110410

BibTeX
@article{Li2025,
  author = {Li, Long and Jia, Tinglei and Yue, Huaizhi and Cheng, Huize and Bu, Yongfeng and Zhang, Zhaoyang},
  title = {TASA: Text-Anchored State–Space Alignment for Long-Tailed Image Classification},
  journal = {Journal of Imaging},
  year = {2025},
  volume = {11},
  number = {11},
  pages = {410},
  doi = {10.3390/jimaging11110410}
}
Kostenlos starten →