FDP: A Frequency-Decomposition Preprocessing Pipeline for Unsupervised Anomaly Detection in Brain MRI

2025-11-16

AI Summary

In unsupervised anomaly detection (UAD) for brain MRI, synthetic perturbations used to simulate lesions often lack pathological fidelity, limiting clinical reliability. Method: This paper proposes the Frequency Decomposition Preprocessing (FDP) framework—a plug-and-play module that leverages Fourier-domain analysis to reveal discriminative spectral signatures of anomalies. FDP implements a dual-path reconstruction mechanism: low-frequency components, which exhibit stability in healthy tissue, are exploited to suppress pathological signals while preserving anatomical structure. Contribution/Results: FDP is model-agnostic and seamlessly integrates with generative models. When applied to Latent Diffusion Models (LDMs), it improves Dice score by 17.63%, significantly enhancing both anomaly localization accuracy and diagnostic fidelity. By establishing a frequency-domain modeling paradigm grounded in biologically interpretable spectral characteristics, FDP advances UAD toward greater clinical credibility.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization

Abstract

Due to the diversity of brain anatomy and the scarcity of annotated data, supervised anomaly detection for brain MRI remains challenging, driving the development of unsupervised anomaly detection (UAD) approaches. Current UAD methods typically utilize artificially generated noise perturbations on healthy MRIs to train generative models for normal anatomy reconstruction, enabling anomaly detection via residual mapping. However, such simulated anomalies lack the biophysical fidelity and morphological complexity characteristic of true clinical lesions. To advance UAD in brain MRI, we conduct the first systematic frequency-domain analysis of pathological signatures, revealing two key properties: (1) anomalies exhibit unique frequency patterns distinguishable from normal anatomy, and (2) low-frequency signals maintain consistent representations across healthy scans. These insights motivate our Frequency-Decomposition Preprocessing (FDP) framework, the first UAD method to leverage frequency-domain reconstruction for simultaneous pathology suppression and anatomical preservation. FDP can integrate seamlessly with existing anomaly simulation techniques, consistently enhancing detection performance across diverse architectures while maintaining diagnostic fidelity. Experimental results demonstrate that FDP consistently improves anomaly detection performance when integrated with existing methods. Notably, FDP achieves a 17.63% increase in DICE score with LDM while maintaining robust improvements across multiple baselines. The code is available at https://github.com/ls1rius/MRI_FDP.

Problem

• Addressing unrealistic simulated anomalies in unsupervised brain MRI detection
• Developing frequency-domain analysis to distinguish pathological from normal anatomy
• Enhancing anomaly detection performance while preserving diagnostic fidelity

Innovation

• Frequency-decomposition preprocessing for anomaly detection
• Leverages frequency-domain reconstruction to suppress pathology
• Enhances existing methods with consistent performance improvements