KFS$^4$LAM: Adaptive KeyFrame Selection with Structure-Aware Mapping for Endoscopic Gaussian Splatting SLAM

Key points

• Gaussian-splatting Simultaneous Localization and Mapping (SLAM) augments conventional SLAM with a dense, renderable 3D scene representation, enabling high-quality reconstruction and novel-view synthesis.1
• Accumulated motion guides adaptive keyframe selection, with complementary cues suppressing redundant or unreliable candidates.1
• Density-aware mapping with anatomical contour supervision (colon folds) emphasizes structurally informative regions, while restricted temporal deformation accommodates local tissue dynamics without unconstrained geometric drift.1
• We evaluate KFS$^4$LAM against Gaussian-SLAM baselines on C3VD and C3VD-v2, together with cross-backbone and component ablations.1

Sources (1)

• [1]KFS$^4$LAM: Adaptive KeyFrame Selection with Structure-Aware Mapping for Endoscopic Gaussian Splatting SLAM
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 15, 06:02 PM

  We present KFS$^4$LAM (KeyFrame Selection with Structure-aware mapping for endoscopic Gaussian Splatting SLAM), which couples motion-guided adaptive keyframe selection with structurally supervised mapping and restricted temporal deformation.

  Gaussian-splatting Simultaneous Localization and Mapping (SLAM) augments conventional SLAM with a dense, renderable 3D scene representation, enabling high-quality reconstruction and novel-view synthesis.

Extractive summary: sentences quoted from the sources.