[ICCV 25 Highlight] MetaScope: Optics-Driven Neural Network for Ultra-Micro Metalens Endoscopy

[Project Page] [Paper Link]

📌 This is the official PyTorch implementation of the work (Full score in final rating: 6,6,6):

MetaScope: Optics-Driven Neural Network for Ultra-Micro Metalens Endoscopy
Wuyang Li ^1,3,*, Wentao Pan^1,*, Xiaoyuan Liu^2,3,*, Zhendong Luo², Chenxin Li¹, Hengyu Liu¹, Din Ping Tsai^2,†, Mu Ku Chen^2,†, Yixuan Yuan^1,†
1 The Chinese University of Hong Kong; ² City University of Hong Kong; ³ École Polytechnique Fédérale de Lausanne (EPFL)
^* Equal Controbution; ^† Corresponding Authors.

Everything is coming soon.

Contact: wuyang.li@epfl.ch

✨ Highlight

We pioneer a new direction in advancing in-vivo intelligence by integrating THREE TYPES OF SCIENCE: optical science, biological science, and computer science.

🎯 Takehome message

• Paper Motivation: How to initiate a new AI-for-Science direction by drawing inspiration from interdisciplinary fields.
• Methodology Design: Approaches for developing methods inspired by nature and scientific principles.
• Academic Path: Our benchmark (requires only 2 commercial GPUs) is easy to follow, which may help you publish follow-up innovative papers and support academic journey, such as graduation.
• Experimental Validation: Strategies for designing tailored experiments to rigorously justify each module.

📋 TODO List

• Release the arXiv paper
• Release the benchmark datasets
• Release the unified codebase
• Release all models

📚Citeation

If you think our work is helpful for your project, I would greatly appreciate it if you could consider citing our work.

@article{li2025metascope,
  title={MetaScope: Optics-Driven Neural Network for Ultra-Micro Metalens Endoscopy},
  author={Li, Wuyang and Pan, Wentao and Liu, Xiaoyuan and Luo, Zhendong and Li, Chenxin and Liu, Hengyu and Tsai, Din Ping and Chen, Mu Ku and Yuan, Yixuan},
  journal={arXiv preprint arXiv:2508.03596},
  year={2025}
}