EndoSfM3D: Learning to 3D Reconstruct Any Endoscopic Surgery Scene using Self-supervised Foundation Model

TL;DR

This paper presents EndoSfM3D, a self-supervised framework that jointly estimates depth, pose, and intrinsic parameters for endoscopic scenes, improving 3D reconstruction accuracy in challenging surgical environments.

Contribution

It introduces a novel joint estimation method integrating intrinsic calibration into self-supervised monocular depth estimation for endoscopy.

Findings

Outperforms state-of-the-art in depth and 3D reconstruction accuracy

Effectively estimates intrinsic parameters in real surgical settings

Validated on public datasets with superior results

Abstract

3D reconstruction of endoscopic surgery scenes plays a vital role in enhancing scene perception, enabling AR visualization, and supporting context-aware decision-making in image-guided surgery. A critical yet challenging step in this process is the accurate estimation of the endoscope's intrinsic parameters. In real surgical settings, intrinsic calibration is hindered by sterility constraints and the use of specialized endoscopes with continuous zoom and telescope rotation. Most existing methods for endoscopic 3D reconstruction do not estimate intrinsic parameters, limiting their effectiveness for accurate and reliable reconstruction. In this paper, we integrate intrinsic parameter estimation into a self-supervised monocular depth estimation framework by adapting the Depth Anything V2 (DA2) model for joint depth, pose, and intrinsics prediction. We introduce an attention-based pose network and a Weight-Decomposed Low-Rank Adaptation (DoRA) strategy for efficient fine-tuning of DA2. Our method is validated on the SCARED and C3VD public datasets, demonstrating superior performance compared to recent state-of-the-art approaches in self-supervised monocular depth estimation and 3D reconstruction. Code and model weights can be found in project repository: https://github.com/MOYF-beta/EndoSfM3D.