Unsupervised 3D out-of-distribution detection with latent diffusion models

TL;DR

This paper introduces a novel method using Latent Diffusion Models for effective out-of-distribution detection in high-resolution 3D medical data, outperforming existing approaches in accuracy and efficiency.

Contribution

The work adapts Latent Diffusion Models for 3D OOD detection, demonstrating improved performance and scalability over previous 2D and 3D methods.

Findings

LDM-based approach achieves statistically significant better performance.

LDM shows less sensitivity to latent representation variations.

LDM provides better spatial anomaly maps.

Abstract

Methods for out-of-distribution (OOD) detection that scale to 3D data are crucial components of any real-world clinical deep learning system. Classic denoising diffusion probabilistic models (DDPMs) have been recently proposed as a robust way to perform reconstruction-based OOD detection on 2D datasets, but do not trivially scale to 3D data. In this work, we propose to use Latent Diffusion Models (LDMs), which enable the scaling of DDPMs to high-resolution 3D medical data. We validate the proposed approach on near- and far-OOD datasets and compare it to a recently proposed, 3D-enabled approach using Latent Transformer Models (LTMs). Not only does the proposed LDM-based approach achieve statistically significant better performance, it also shows less sensitivity to the underlying latent representation, more favourable memory scaling, and produces better spatial anomaly maps. Code is available at https://github.com/marksgraham/ddpm-ood