Reproduce “Semantic-SuPer: A Semantic-aware Surgical Perception Framework for Endoscopic Tissue Identification, Reconstruction, and Tracking” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Shan Lin, Albert J. Miao, Jingpei Lu, Shunkai Yu, Zih-Yun Chiu, Florian Richter, Michael C. Yip
The Semantic-SuPer framework integrates pixel-wise semantic labels from endoscopic video with 3D deformation tracking and reconstruction, enhancing perception for identifying, reconstructing, and tracking deformable surgical tissues. This approach achieves lower mean reprojection errors, for example, 7.5 pixels compared to 10.8 pixels for the baseline, particularly during significant tissue deformation.
Key insights
- Integrating real-time 2D semantic information directly from endoscopic video significantly improves the robustness and accuracy of 3D deformable tissue tracking and reconstruction in complex surgical scenes.
- The proposed semantic-aware morphing loss effectively guides deformation optimization by explicitly encouraging semantic edge consistency, which is crucial for handling large deformations and noisy depth estimates.
- The accuracy of the underlying semantic segmentation directly impacts the overall performance of the integrated 3D surgical perception framework, highlighting the importance of high-quality semantic inputs.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.