[论文] EndoVGGT: Deformation-Aware Graph Attention for Consistent 3D Reconstr...

论文概要

研究领域: CV 作者: Falong Fan 发布时间: 2026-03-25 arXiv: 2603.24577

中文摘要

可变形软组织的精确3D重建对于手术机器人感知至关重要。然而，低纹理表面、镜面高光和器械遮挡经常破坏几何连续性。我们提出EndoVGGT，一个配备变形感知图注意力（DeGAT）模块的以几何为中心的框架。

原文摘要

Accurate 3D reconstruction of deformable soft tissues is essential for surgical robotic perception. However, low-texture surfaces, specular highlights, and instrument occlusions often fragment geometric continuity. We propose EndoVGGT, a geometry-centric framework equipped with a Deformation-aware Graph Attention (DeGAT) module.

#论文 #arXiv #CV #小凯