Unlocking Dark Vision Potential for Medical Image Segmentation

Yang, Hongpeng; Hu, Xiangyu; Chen, Yingxin; Chen, Siyu; Nelakuditi, Srihari; Tong, Yan; Ma, Shiqiang; Guo, Fei

doi:10.24963/ijcai.2025/245

Unlocking Dark Vision Potential for Medical Image Segmentation

Hongpeng Yang, Xiangyu Hu, Yingxin Chen, Siyu Chen, Srihari Nelakuditi, Yan Tong, Shiqiang Ma, Fei Guo

Download BibTeX Cite

Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence. Main Track. Pages 2197-2205.

https://doi.org/10.24963/ijcai.2025/245

Keywords

Computer Vision: CV: Scene analysis and understanding
Computer Vision: CV: Biomedical image analysis
Natural Language Processing: General
Computer Vision: CV: Machine learning for vision

Abstract

Accurate segmentation of lesions is crucial for disease diagnosis and treatment planning. However, blurring and low contrast in the imaging process can affect segmentation results. We have observed that noninvasive medical imaging shares considerable similarities with natural images under low light conditions and that nocturnal animals possess extremely strong night vision capabilities. Inspired by the dark vision of these nocturnal animals, we proposed a novel plug-and-play dark vision network (DVNet) to enhance the model's perception for low-contrast medical images. Specifically, by employing the wavelet transform, we decompose medical images into subbands of varying frequencies, mimicking the sensitivity of photoreceptor cells to different light intensities. To simulate the antagonistic receptive fields of horizontal cells and bipolar cells, we design a Mamba-Enhanced Fusion Module to achieve global information correlation and enhance contrast between lesions and surrounding healthy tissues. Extensive experiments demonstrate that the DVNet achieves SOTA performance in various medical image segmentation tasks.

Hongpeng Yang, Xiangyu Hu, Yingxin Chen, Siyu Chen, Srihari Nelakuditi, Yan Tong, Shiqiang Ma, and Fei Guo. 2025. Unlocking Dark Vision Potential for Medical Image Segmentation. In Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence, IJCAI-25. IJCAI, 2197–2205. https://doi.org/10.24963/ijcai.2025/245

@inproceedings{ijcai2025p245,
  title     = {Unlocking Dark Vision Potential for Medical Image Segmentation},
  author    = {Yang, Hongpeng and Hu, Xiangyu and Chen, Yingxin and Chen, Siyu and Nelakuditi, Srihari and Tong, Yan and Ma, Shiqiang and Guo, Fei},
  booktitle = {Proceedings of the Thirty-Fourth International Joint Conference on
               Artificial Intelligence, {IJCAI-25}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},
  editor    = {James Kwok},
  pages     = {2197--2205},
  year      = {2025},
  month     = {8},
  note      = {Main Track},
  doi       = {10.24963/ijcai.2025/245},
  url       = {https://doi.org/10.24963/ijcai.2025/245},
}

TY  - CPAPER
TI  - Unlocking Dark Vision Potential for Medical Image Segmentation
AU  - Yang, Hongpeng
AU  - Hu, Xiangyu
AU  - Chen, Yingxin
AU  - Chen, Siyu
AU  - Nelakuditi, Srihari
AU  - Tong, Yan
AU  - Ma, Shiqiang
AU  - Guo, Fei
T2  - Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence, IJCAI-25
PB  - International Joint Conferences on Artificial Intelligence Organization
PY  - 2025
SP  - 2197
EP  - 2205
DO  - 10.24963/ijcai.2025/245
UR  - https://doi.org/10.24963/ijcai.2025/245
ER  -