Reproduce “Tracking monocular camera pose and deformation for SLAM inside the human body” 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
Juan J. Gomez Rodriguez, J.M.M Montiel, Juan D. Tardos
Researchers at the Universidad de Zaragoza developed the first pure monocular method to simultaneously track camera pose and 3D scene deformation within the human body. This approach operates in general, unstructured scenes, overcoming prior topological assumptions and demonstrating robust performance on both simulated and real colonoscopy data.
Key insights
- By assuming camera motion is significantly faster than deformation, the system can reliably initialize a 3D map from monocular frames and effectively separate camera motion from scene deformation.
- Adaptive photometric feature tracking with local illumination gain/bias and SSIM-based outlier rejection provides robust data association in challenging endoscopic lighting and texture conditions.
- Combining spatial (as-rigid-as-possible) and temporal deformation regularizers within a unified optimization framework allows for the tracking of complex, unstructured tissue deformations without relying on restrictive global models.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.