Monocular pose estimation of articulated open surgery tools -- in the wild

TL;DR

This paper introduces a novel monocular 6D pose estimation framework for articulated surgical tools in open surgery, leveraging synthetic data, domain adaptation, and detection techniques to operate effectively in real-world scenarios.

Contribution

It presents a comprehensive framework combining synthetic data generation, pose estimation, and domain adaptation for surgical tools, reducing manual annotation needs.

Findings

Effective pose estimation on real surgical data

Robust to occlusions and specularity

Potential for integration into medical AR and robotic systems

Abstract

This work presents a framework for monocular 6D pose estimation of surgical instruments in open surgery, addressing challenges such as object articulations, specularity, occlusions, and synthetic-to-real domain adaptation. The proposed approach consists of three main components: $(1)$ synthetic data generation pipeline that incorporates 3D scanning of surgical tools with articulation rigging and physically-based rendering; $(2)$ a tailored pose estimation framework combining tool detection with pose and articulation estimation; and $(3)$ a training strategy on synthetic and real unannotated video data, employing domain adaptation with automatically generated pseudo-labels. Evaluations conducted on real data of open surgery demonstrate the good performance and real-world applicability of the proposed framework, highlighting its potential for integration into medical augmented reality and robotic systems. The approach eliminates the need for extensive manual annotation of real surgical data.