PhySe-RPO: Physics and Semantics Guided Relative Policy Optimization for Diffusion-Based Surgical Smoke Removal

TL;DR

PhySe-RPO introduces a physics and semantics-guided diffusion framework for surgical smoke removal, enabling robust, interpretable restoration with limited supervision in surgical videos.

Contribution

It transforms deterministic restoration into a stochastic policy with physics and semantic rewards, improving robustness and interpretability in surgical smoke removal.

Findings

Produces physically consistent and semantically faithful restorations.

Effective on both synthetic and real surgical datasets.

Operates without requiring paired supervision.

Abstract

Surgical smoke severely degrades intraoperative video quality, obscuring anatomical structures and limiting surgical perception. Existing learning-based desmoking approaches rely on scarce paired supervision and deterministic restoration pipelines, making it difficult to perform exploration or reinforcement-driven refinement under real surgical conditions. We propose PhySe-RPO, a diffusion restoration framework optimized through Physics- and Semantics-Guided Relative Policy Optimization. The core idea is to transform deterministic restoration into a stochastic policy, enabling trajectory-level exploration and critic-free updates via group-relative optimization. A physics-guided reward imposes illumination and color consistency, while a visual-concept semantic reward learned from CLIP-based surgical concepts promotes smoke-free and anatomically coherent restoration. Together with a reference-free perceptual constraint, PhySe-RPO produces results that are physically consistent, semantically faithful, and clinically interpretable across synthetic and real robotic surgical datasets, providing a principled route to robust diffusion-based restoration under limited paired supervision.