IterMask3D: Unsupervised Anomaly Detection and Segmentation with Test-Time Iterative Mask Refinement in 3D Brain MR

2025-04-07

AI Summary

Existing unsupervised 3D brain MRI anomaly detection methods rely on image corruption during training to model the “normal” distribution, but this global degradation strategy discards normal-region information and exacerbates reconstruction artifacts, leading to elevated false positives. To address this, we propose Test-time Iterative Spatial Mask Optimization (TISMO): instead of applying uniform degradation, TISMO adaptively refines a spatial mask at test time via error-driven iterative contraction, progressively unmasking normal regions while preserving structural integrity. Furthermore, high-frequency structural guidance is incorporated into reconstruction to enhance fidelity. To our knowledge, TISMO is the first unsupervised framework to perform adaptive spatial mask optimization exclusively during inference. Evaluated on multi-sequence MRI data—including scans with synthetic artifacts and real clinical pathologies (tumors, strokes)—TISMO achieves state-of-the-art performance: reducing false positive rate by 18.7% and improving Dice score by 5.2%.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Multimodal Learning
• Search and Optimization: Non-convex Optimization

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Unsupervised anomaly detection and segmentation methods train a model to learn the training distribution as 'normal'. In the testing phase, they identify patterns that deviate from this normal distribution as 'anomalies'. To learn the `normal' distribution, prevailing methods corrupt the images and train a model to reconstruct them. During testing, the model attempts to reconstruct corrupted inputs based on the learned 'normal' distribution. Deviations from this distribution lead to high reconstruction errors, which indicate potential anomalies. However, corrupting an input image inevitably causes information loss even in normal regions, leading to suboptimal reconstruction and an increased risk of false positives. To alleviate this, we propose IterMask3D, an iterative spatial mask-refining strategy designed for 3D brain MRI. We iteratively spatially mask areas of the image as corruption and reconstruct them, then shrink the mask based on reconstruction error. This process iteratively unmasks 'normal' areas to the model, whose information further guides reconstruction of 'normal' patterns under the mask to be reconstructed accurately, reducing false positives. In addition, to achieve better reconstruction performance, we also propose using high-frequency image content as additional structural information to guide the reconstruction of the masked area. Extensive experiments on the detection of both synthetic and real-world imaging artifacts, as well as segmentation of various pathological lesions across multiple MRI sequences, consistently demonstrate the effectiveness of our proposed method.

Problem

• Detects anomalies in 3D brain MRI unsupervised
• Reduces false positives via iterative mask refinement
• Improves reconstruction using high-frequency image guidance

Innovation

• Iterative spatial mask-refining strategy for 3D MRI
• High-frequency image content guides reconstruction
• Reduces false positives in anomaly detection