arXiv cs.AI· Anna Oliveras, Roger Mar\'i, Rafael Redondo, Oriol Guardi\`a, Cynthia Ifeyinwa Ugwu, Ana Tost, Bhalaji Nagarajan, Carolina Migliorelli, Vicent Ribas, Petia Radeva·· 13 小时前AI 评分20
MUMINS：元数据条件化的不确定性感知医学图像下一状态合成
MUMINS: Metadata-conditioned Uncertainty-aware Medical Image Next-state Synthesis
AI 导读
MUMINS 是一个高效扩散框架，在单次逆向扩散过程中同时扩散基线扫描与随访残差并合成随访图像，同步预测空间不确定性图。该模型以时间间隔和元数据为条件，在每个去噪步骤动态回注基线作为软锚点，并用负对数似然头学习不确定性图以标记易错区域。在肺 CT（PNG）和脑 MRI（OASIS-3）上，按数据集分别重训练的 MUMINS 达到或超过专用领域 SOTA 方法。
正文 · 原文
Abstract:Forecasting anatomical changes such as tumor growth and neurodegeneration is a challenging generative vision task. Morphological evolution is subtle relative to static anatomy, highly patient-specific, and inherently stochastic. Existing methods struggle with several issues: deterministic networks ignore biological stochasticity, while standard diffusion models require computationally prohibitive multi-pass sampling to quantify uncertainty. We propose MUMINS (Metadata-conditioned Uncertainty-aware Medical Image Next-state Synthesis), an efficient diffusion framework that jointly diffuses a baseline scan and its follow-up residual, summed to synthesize the follow-up scan, while concurrently predicting a spatial uncertainty map, in a single reverse diffusion process. Conditioned on the time interval and relevant metadata, it preserves fine-grained anatomy by dynamically re-injecting the baseline as a soft anchor at every denoising step, and a negative-log-likelihood head learns the uncertainty map to explicitly flag error-prone regions. Designed without organ-specific heuristics, the same architecture is reused across anatomies via separate, dataset-specific retraining. Extensive evaluations demonstrate that dataset-specific retraining of MUMINS matches or outperforms dedicated, domain-specific state-of-the-art methods on lung CT (PNG) and brain MRI (OASIS-3). Project page: this https URL.
| Comments: | Supplementary material to follow in future versions |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| ACM classes: | I.2.6; I.4.5; I.4.9; J.3 |
| Cite as: | arXiv:2609.17169 [cs.CV] |
| (or arXiv:2609.17169v2 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2609.17169 arXiv-issued DOI via DataCite |
Submission history
From: Anna Oliveras [view email]
[v1]
Tue, 15 Sep 2026 13:32:37 UTC (1,538 KB)
[v2]
Tue, 27 Oct 2026 11:31:01 UTC (1,464 KB)
来源：arXiv cs.AI · arxiv.org