Steerable Conditional Diffusion for Out-of-Distribution Adaptation in Imaging Inverse Problems
Steerable Conditional Diffusion for Out-of-Distribution Adaptation in Imaging Inverse Problems
复制标题
DOI：
10.48550/arxiv.2308.14409
复制
发表时间：
2023-08
期刊：
影响因子：
--
通讯作者：
Riccardo Barbano;Alexander Denker;Hyungjin Chung;Tae Hoon Roh;Simon Arrdige;P. Maass;Bangti Jin;J. C. Ye
中科院分区：
文献类型：
--
作者：
Riccardo Barbano;Alexander Denker;Hyungjin Chung;Tae Hoon Roh;Simon Arrdige;P. Maass;Bangti Jin;J. C. Ye
去噪扩散模型已成为解决成像逆问题的首选框架。关于这些模型的一个关键问题是它们在分布外（OOD）任务上的性能，这仍然是一个未被充分探索的挑战。可以生成与测量数据不一致的真实重建，产生训练数据集中唯一存在的幻觉图像特征。为了同时加强数据一致性和利用数据驱动的先验，我们引入了一种新的采样框架，称为可控条件扩散。该框架使去噪网络专门适应于可用的测量数据。利用我们提出的方法，我们在不同成像模式下实现了OOD性能的实质性增强，推进了去噪扩散模型在现实应用中的鲁棒部署。
Denoising diffusion models have emerged as the go-to framework for solving inverse problems in imaging. A critical concern regarding these models is their performance on out-of-distribution (OOD) tasks, which remains an under-explored challenge. Realistic reconstructions inconsistent with the measured data can be generated, hallucinating image features that are uniquely present in the training dataset. To simultaneously enforce data-consistency and leverage data-driven priors, we introduce a novel sampling framework called Steerable Conditional Diffusion. This framework adapts the denoising network specifically to the available measured data. Utilising our proposed method, we achieve substantial enhancements in OOD performance across diverse imaging modalities, advancing the robust deployment of denoising diffusion models in real-world applications.