arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2605.10076eess.IVcs.LG
收藏

生成正则化器在反问题中的稳定性基准

A Stability Benchmark of Generative Regularizers for Inverse Problems

Helmholtz Imaging, Deutsches Elektronen-Synchrotron DESY, Germany(德意志电子同步辐射研究中心(Helmholtz Imaging));ENS Paris & Inria, France(巴黎高等师范学院与法国国家信息与自动化研究所);Chemnitz University of Technology, Germany(化学工业大学)

Alexander Denker, Johannes Hertrich, Sebastian Neumayer

AI总结:

本文评估生成正则化器在反问题中的稳定性,对比优化方法,探讨其在不同场景下的性能。

摘要:

生成(扩散)先验在成像反问题中表现出色,但科学和医学成像需确保在不完美设置下重建技术稳定可靠。典型稳定性定义包括''收敛正则化''、对分布外数据的鲁棒性以及对前向算子或噪声模型不准确性的鲁棒性。我们通过数值方法评估这些属性。此外,我们将生成方法与受广泛使用变分技术启发的现代优化方法进行对比。我们的结果为生成先验在哪些设置和应用中能提供最先进的重建,以及在哪些情况下表现不足或可能存在问题提供了见解。

英文摘要:

Generative (diffusion) priors demonstrate remarkable performance in addressing inverse problems in imaging. Yet, for scientific and medical imaging, it is crucial that reconstruction techniques remain stable and reliable under imperfect settings. Typical definitions of stability encompass the notion of ''convergent regularization'', robustness to out-of-distribution data, and to inaccuracies in the forward operator or noise model. We evaluate these properties numerically. Furthermore, we benchmark generative approaches against modern optimization-based methods inspired by the widely used variational techniques. Our results give insights for which settings and applications generative priors can deliver state-of-the-art reconstructions, and on those in which they fall short or may even be problematic.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑