Unsupervised Anomaly Detection (UAD) methods rely on healthy data distributions to identify anomalies as outliers. In brain MRI, a common approach is reconstruction-based UAD, where generative models reconstruct healthy brain MRIs, and anomalies are detected as deviations between input and reconstruction. However, this method is sensitive to imperfect reconstructions, leading to false positives that impede the segmentation. To address this limitation, we construct multiple reconstructions with probabilistic diffusion models. We then analyze the resulting distribution of these reconstructions using the Mahalanobis distance to identify anomalies as outliers. By leveraging information about normal variations and covariance of individual pixels within this distribution, we effectively refine anomaly scoring, leading to improved segmentation. Our experimental results demonstrate substantial performance improvements across various data sets. Specifically, compared to relying solely on single reconstructions, our approach achieves relative improvements of 15.9%, 35.4%, 48.0%, and 4.7% in terms of AUPRC for the BRATS21, ATLAS, MSLUB and WMH data sets, respectively.

翻译：无监督异常检测方法依赖健康数据分布，将异常识别为离群值。在脑部MRI中，一种常见方法是基于重建的无监督异常检测，即生成模型重建健康脑部MRI，并通过输入与重建之间的偏差来检测异常。然而，该方法对不完美重建较为敏感，易产生假阳性结果，从而阻碍分割性能。为克服这一局限，我们采用概率扩散模型生成多重重建。随后，利用马氏距离分析这些重建结果的分布，从而将异常识别为统计离群点。通过有效利用该分布中正常变异及像素间协方差信息，我们显著优化了异常评分机制，进而提升了分割精度。实验结果表明，该方法在多个数据集上均取得显著性能提升。具体而言，相较于仅依赖单一重建的方法，本方法在BRATS21、ATLAS、MSLUB和WMH数据集上的AUPRC指标分别实现了15.9%、35.4%、48.0%和4.7%的相对提升。