EUVIMEDEUVIMED entdecken
Crossref

Hi-Mamba: Hierarchical Mamba for Efficient Image Super-Resolution

Junbo Qiao, Jincheng Liao, Wei Li, Yulun Zhang, Yong Guo, Jiao Xie, Jie Hu, Shaohui Lin

IEEE Transactions on Image Processing · 2025 · 34 · 8461-8473

DOI: 10.1109/tip.2025.3643146 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Image Processing, 2025. Verfasst von Junbo Qiao und 7 weiteren Autor:innen. Die Arbeit wurde bisher 16-mal zitiert (Crossref) und nennt 75 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Shanghai Science and Technology Commission, Open Research Fund of the Key Laboratory of Advanced Theory and Application in Statistics and Data Science, Ministry of Education.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Despite Transformers have achieved significant success in low-level vision tasks, they are constrained by computing self-attention with a quadratic complexity and limited-size windows. This limitation results in a lack of global receptive field across the entire image. Recently, State Space Models (SSMs) have gained widespread attention due to their global receptive field and linear complexity with respect to input length. However, integrating SSMs into low-level vision tasks presents two major challenges: 1) Relationship degradation of long-range tokens with a long-range forgetting problem by encoding pixel-by-pixel high-resolution images. 2) Significant redundancy in the existing multi-direction scanning strategy. To this end, we propose Hi-Mamba for image super-resolution (SR) to address these challenges, which unfolds the image with only a single scan. Specifically, the Global Hierarchical Mamba Block (GHMB) enables token interactions across the entire image, providing a global receptive field while leveraging a multi-scale structure to facilitate long-range dependency learning. Additionally, the Direction Alternation Module (DAM) adjusts the scanning patterns of GHMB across different layers to enhance spatial relationship modeling. Extensive experiments demonstrate that our Hi-Mamba achieves 0.2-0.27dB PSNR gains on the Urban100 dataset across different scaling factors compared to the state-of-the-art MambaIRv2 for SR. Moreover, our lightweight Hi-Mamba also outperforms lightweight SRFormer by 0.39dB PSNR for $\times 2$ SR.

Förderung

  • National Natural Science Foundation of China
  • Shanghai Science and Technology Commission
  • Open Research Fund of the Key Laboratory of Advanced Theory and Application in Statistics and Data Science, Ministry of Education
  • Fundamental Research Funds for the Central Universities

Zitieren

APA

Qiao, J., Liao, J., Li, W., Zhang, Y., Guo, Y., Xie, J., Hu, J., & Lin, S. (2025). Hi-Mamba: Hierarchical Mamba for Efficient Image Super-Resolution. IEEE Transactions on Image Processing, 34, 8461-8473. https://doi.org/10.1109/tip.2025.3643146

Vancouver

Qiao J, Liao J, Li W, Zhang Y, Guo Y, Xie J, et al. Hi-Mamba: Hierarchical Mamba for Efficient Image Super-Resolution. IEEE Transactions on Image Processing. 2025;34:8461-8473. doi:10.1109/tip.2025.3643146

BibTeX
@article{Qiao2025,
  author = {Qiao, Junbo and Liao, Jincheng and Li, Wei and Zhang, Yulun and Guo, Yong and Xie, Jiao and Hu, Jie and Lin, Shaohui},
  title = {Hi-Mamba: Hierarchical Mamba for Efficient Image Super-Resolution},
  journal = {IEEE Transactions on Image Processing},
  year = {2025},
  volume = {34},
  pages = {8461--8473},
  doi = {10.1109/tip.2025.3643146}
}
Kostenlos starten →