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

arXiv: https://arxiv.org/abs/2603.22844 · HTML: https://mlsanity.com/paper/2603.22844/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 30 of its 48 references that are in the map.
- First posted on arXiv 24 Mar 2026.

## Builds on (most cited first, 30 of 30)

- [Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2006.11239/) (2020)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Score-Based Generative Modeling through Stochastic Differential Equations](https://mlsanity.com/paper/2011.13456/) (2020)
- [Improved Techniques for Training GANs](https://mlsanity.com/paper/1606.03498/) (2016)
- [Learning to Prompt for Vision-Language Models](https://mlsanity.com/paper/2109.01134/) (2021)
- [MUSIQ: Multi-scale Image Quality Transformer](https://mlsanity.com/paper/2108.05997/) (2021)
- [The Perception-Distortion Tradeoff](https://mlsanity.com/paper/1711.06077/) (2017)
- [MANIQA: Multi-dimension Attention Network for No-Reference Image Quality Assessment](https://mlsanity.com/paper/2204.08958/) (2022)
- [Blind Image Quality Assessment via Vision-Language Correspondence: A Multitask Learning Perspective](https://mlsanity.com/paper/2303.14968/) (2023)
- [Vision Transformers for Single Image Dehazing](https://mlsanity.com/paper/2204.03883/) (2022)
- [Improving Vision-Language-Action Model with Online Reinforcement Learning](https://mlsanity.com/paper/2501.16664/) (2025)
- [Controlling Vision-Language Models for Multi-Task Image Restoration](https://mlsanity.com/paper/2310.01018/) (2023)
- [SimpleAR: Pushing the Frontier of Autoregressive Visual Generation through Pretraining, SFT, and RL](https://mlsanity.com/paper/2504.11455/) (2025)
- [Seg-R1: Segmentation Can Be Surprisingly Simple with Reinforcement Learning](https://mlsanity.com/paper/2506.22624/) (2025)
- [Structure Representation Network and Uncertainty Feedback Learning for Dense Non-Uniform Fog Removal](https://mlsanity.com/paper/2210.03061/) (2022)
- [MixDehazeNet : Mix Structure Block For Image Dehazing Network](https://mlsanity.com/paper/2305.17654/) (2023)
- [Reasoning Physical Video Generation with Diffusion Timestep Tokens via Reinforcement Learning](https://mlsanity.com/paper/2504.15932/) (2025)
- [No-Reference Quality Assessment of Contrast-Distorted Images using Contrast Enhancement](https://mlsanity.com/paper/1904.08879/) (2019)
- [LighTDiff: Surgical Endoscopic Image Low-Light Enhancement with T-Diffusion](https://mlsanity.com/paper/2405.10550/) (2024)
- [Self-Supervised Video Desmoking for Laparoscopic Surgery](https://mlsanity.com/paper/2403.11192/) (2024)
- [Desmoking laparoscopy surgery images using an image-to-image translation guided by an embedded dark channel](https://mlsanity.com/paper/2004.08947/) (2020)
- [Denoising as Adaptation: Noise-Space Domain Adaptation for Image Restoration](https://mlsanity.com/paper/2406.18516/) (2024)
- [Towards Self-Improvement of Diffusion Models via Group Preference Optimization](https://mlsanity.com/paper/2505.11070/) (2025)
- [Temporal As a Plugin: Unsupervised Video Denoising with Pre-Trained Image Denoisers](https://mlsanity.com/paper/2409.11256/) (2024)
- [Self-Reflective Reinforcement Learning for Diffusion-based Image Reasoning Generation](https://mlsanity.com/paper/2505.22407/) (2025)
- [Generative Smoke Removal](https://mlsanity.com/paper/1902.00311/) (2019)
- [Progressive Frequency-Aware Network for Laparoscopic Image Desmoking](https://mlsanity.com/paper/2312.12023/) (2023)
- [LSD3K: A Benchmark for Smoke Removal from Laparoscopic Surgery Images](https://mlsanity.com/paper/2407.13132/) (2024)
- [AMNCutter: Affinity-Attention-Guided Multi-View Normalized Cutter for Unsupervised Surgical Instrument Segmentation](https://mlsanity.com/paper/2411.03695/) (2024)
- [Efficient Dual-domain Image Dehazing with Haze Prior Perception](https://mlsanity.com/paper/2507.11035/) (2025)

## Abstract (arXiv)

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.