Reproduce “Endo-TSR: Temporal Spectral Modeling of Appearance and Motion for Endoscopic Reconstruction” 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
Taoyu Wu, Yiyi Miao, Qi Shao, Zhuoxiao Li, Zhe Tang, Limin Yu, Baoru Huang
Endo-TSR reconstructs deforming endoscopic scenes by modeling each Gaussian’s colour and translation with separate time-varying corrections. This lets the representation adjust appearance independently of position. In matched-input comparisons with SurgicalGS, it improves PSNR by 0.31–0.66 dB and reduces LPIPS by 0.0035–0.0096. Across six sequences, it has highest PSNR, but Endo-PairGS has lower LPIPS on Pulling. The results measure held-out rendering.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.