Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2603.06173 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910043629158400 |
|---|---|
| author | Tian, Yueying Han, Xudong Zhou, Meng Aviles-Espinosa, Rodrigo Young, Rupert Birch, Philip |
| author_facet | Tian, Yueying Han, Xudong Zhou, Meng Aviles-Espinosa, Rodrigo Young, Rupert Birch, Philip |
| contents | Diffusion models have emerged as powerful tools for 3D medical image generation, yet bridging the gap between standard training objectives and clinical relevance remains a challenge. This paper presents a method to enhance 3D diffusion models using Reinforcement Learning (RL) with multi-scale feedback. We first pretrain a 3D diffusion model on MRI volumes to establish a robust generative prior. Subsequently, we fine-tune the model using Proximal Policy Optimization (PPO), guided by a novel reward system that integrates both 2D slice-wise assessments and 3D volumetric analysis. This combination allows the model to simultaneously optimize for local texture details and global structural coherence. We validate our framework on the BraTS 2019 and OASIS-1 datasets. Our results indicate that incorporating RL feedback effectively steers the generation process toward higher quality distributions. Quantitative analysis reveals significant improvements in Fréchet Inception Distance (FID) and, crucially, the synthetic data demonstrates enhanced utility in downstream tumor and disease classification tasks compared to non-optimized baselines. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_06173 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Optimizing 3D Diffusion Models for Medical Imaging via Multi-Scale Reward Learning Tian, Yueying Han, Xudong Zhou, Meng Aviles-Espinosa, Rodrigo Young, Rupert Birch, Philip Computer Vision and Pattern Recognition Diffusion models have emerged as powerful tools for 3D medical image generation, yet bridging the gap between standard training objectives and clinical relevance remains a challenge. This paper presents a method to enhance 3D diffusion models using Reinforcement Learning (RL) with multi-scale feedback. We first pretrain a 3D diffusion model on MRI volumes to establish a robust generative prior. Subsequently, we fine-tune the model using Proximal Policy Optimization (PPO), guided by a novel reward system that integrates both 2D slice-wise assessments and 3D volumetric analysis. This combination allows the model to simultaneously optimize for local texture details and global structural coherence. We validate our framework on the BraTS 2019 and OASIS-1 datasets. Our results indicate that incorporating RL feedback effectively steers the generation process toward higher quality distributions. Quantitative analysis reveals significant improvements in Fréchet Inception Distance (FID) and, crucially, the synthetic data demonstrates enhanced utility in downstream tumor and disease classification tasks compared to non-optimized baselines. |
| title | Optimizing 3D Diffusion Models for Medical Imaging via Multi-Scale Reward Learning |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2603.06173 |