MAD-AD: Masked Diffusion for Unsupervised Brain Anomaly Detection

2025-02-24

AI Summary

To address the challenges of complex anatomical structures and scarce abnormal annotations in unsupervised anomaly detection for brain MRI, this paper proposes a novel Masked Latent Diffusion Model (MLDM). MLDM operates in the latent space by randomly masking and corrupting image patches with noise, jointly optimizing noise discrimination and feature reconstruction. Anomalies are implicitly modeled as deviations from the learned latent noise distribution, enabling pixel-level anomaly localization and normalizing reconstruction—without requiring any abnormal training samples. Notably, this work is the first to incorporate a masking mechanism into diffusion models to support fine-grained anomaly localization. Evaluated on multiple public brain anomaly datasets, MLDM achieves state-of-the-art performance in both anomaly localization accuracy and reconstruction fidelity, significantly outperforming existing unsupervised methods.

Technology Category

• Machine Learning: Unsupervised & Self-Supervised Learning
• Computer Vision: Diffusion Models for Vision
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• User Modeling, Personalization and Recommendation: ML for personalized search and recommendations
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Unsupervised anomaly detection in brain images is crucial for identifying injuries and pathologies without access to labels. However, the accurate localization of anomalies in medical images remains challenging due to the inherent complexity and variability of brain structures and the scarcity of annotated abnormal data. To address this challenge, we propose a novel approach that incorporates masking within diffusion models, leveraging their generative capabilities to learn robust representations of normal brain anatomy. During training, our model processes only normal brain MRI scans and performs a forward diffusion process in the latent space that adds noise to the features of randomly-selected patches. Following a dual objective, the model learns to identify which patches are noisy and recover their original features. This strategy ensures that the model captures intricate patterns of normal brain structures while isolating potential anomalies as noise in the latent space. At inference, the model identifies noisy patches corresponding to anomalies and generates a normal counterpart for these patches by applying a reverse diffusion process. Our method surpasses existing unsupervised anomaly detection techniques, demonstrating superior performance in generating accurate normal counterparts and localizing anomalies. The code is available at hhttps://github.com/farzad-bz/MAD-AD.

Problem

• Unsupervised anomaly detection in brain images
• Accurate localization of anomalies in medical images
• Leveraging diffusion models for robust brain anatomy representation

Innovation

• Masked Diffusion Models
• Unsupervised Brain Anomaly Detection
• Reverse Diffusion Process