AnoFPDM: Anomaly Segmentation with Forward Process of Diffusion Models for Brain MRI

TL;DR

AnoFPDM introduces a fully weakly-supervised diffusion model framework for brain MRI anomaly segmentation, eliminating the need for pixel-level labels and outperforming recent state-of-the-art methods.

Contribution

It proposes a novel framework that leverages the unguided forward process of diffusion models for hyperparameter tuning without pixel labels.

Findings

Outperforms recent weakly-supervised methods

Operates without pixel-level labels

Enhances anomaly signal detection

Abstract

Weakly-supervised diffusion models (DMs) in anomaly segmentation, leveraging image-level labels, have attracted significant attention for their superior performance compared to unsupervised methods. It eliminates the need for pixel-level labels in training, offering a more cost-effective alternative to supervised methods. However, existing methods are not fully weakly-supervised because they heavily rely on costly pixel-level labels for hyperparameter tuning in inference. To tackle this challenge, we introduce Anomaly Segmentation with Forward Process of Diffusion Models (AnoFPDM), a fully weakly-supervised framework that operates without the need of pixel-level labels. Leveraging the unguided forward process as a reference for the guided forward process, we select hyperparameters such as the noise scale, the threshold for segmentation and the guidance strength. We aggregate anomaly maps from guided forward process, enhancing the signal strength of anomalous regions. Remarkably, our proposed method outperforms recent state-of-the-art weakly-supervised approaches, even without utilizing pixel-level labels.