Problem
Research questions and friction points this paper is trying to address.
Endoscopic scene reconstruction
Deformable Gaussian splatting
Temporal appearance modeling
Tissue motion
Surface detail recovery
Endo-TSR: Temporal Spectral Modeling of Appearance and Motion for Endoscopic Reconstruction
🤖 AI Summary
This study addresses the challenge of simultaneously modeling tissue deformation, temporal appearance variations, and fine surface details in dynamic endoscopic reconstruction. We propose Endo-TSR, a deformable 3D Gaussian Splatting framework that pioneers the decoupled modeling of color and motion under a shared temporal frequency. Specifically, it introduces bounded Fourier color residuals alongside independent translation residuals to achieve effective feature separation, employs a Matérn spectral prior to regularize the motion field, and designs multi-scale Laplacian supervision to recover high-frequency geometric details. Extensive evaluations on the EndoNeRF and StereoMIS datasets demonstrate that Endo-TSR achieves state-of-the-art rendering quality, yielding the highest PSNR across all sequences.
📝 Abstract
Endoscopic scene reconstruction requires modeling tissue motion and temporal appearance while recovering fine surface detail. Deformable Gaussian models provide explicit trajectories, but their fixed colour coefficients lack a dedicated temporal representation for photometric changes. We propose Endo-TSR, which augments deformable Gaussian splatting with bounded Fourier colour residuals and independent translation residuals on shared temporal frequencies. The colour residuals capture local appearance changes, while a Mat\'ern spectral prior regularises motion corrections. Multi-scale Laplacian supervision guides tissue-detail recovery during joint image fitting. Extensive experiments on the EndoNeRF and StereoMIS datasets demonstrate state-of-the-art rendering quality, with the highest PSNR across all evaluated sequences. Ablation studies show that temporal appearance yields the largest PSNR gain among the tested component additions, while appearance and detail supervision jointly improve rendering with fixed Gaussian counts.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Deformable Gaussian Splatting
Temporal Spectral Modeling
Fourier Colour Residuals
Matérn Spectral Prior
Endoscopic Reconstruction
🔎 Similar Papers
No similar papers found.
