Out-of-Distribution Detection in Medical Imaging via Diffusion Trajectories

2025-07-31

AI Summary

This paper addresses unsupervised out-of-distribution (OOD) detection for rare pathologies in medical imaging—characterized by zero annotated labels, severe class imbalance, and distributional shift—where robust and efficient discrimination between normal and abnormal samples is critically challenging. We propose a novel reconstruction-free method grounded in Stein score–driven denoising diffusion models (SBDDM). Our approach extracts curvature features from the forward diffusion trajectory, enabling anomaly scoring within just five diffusion steps. A single pre-trained model generalizes across diverse near-domain and far-domain OOD tasks without retraining. Evaluated on multiple benchmarks, our method achieves state-of-the-art performance, with AUROC improvements of up to 10.43% for near-OOD and 18.10% for far-OOD detection over existing approaches. Moreover, it significantly enhances inference efficiency, offering a scalable, label-free solution for clinical anomaly detection.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling

Abstract

In medical imaging, unsupervised out-of-distribution (OOD) detection offers an attractive approach for identifying pathological cases with extremely low incidence rates. In contrast to supervised methods, OOD-based approaches function without labels and are inherently robust to data imbalances. Current generative approaches often rely on likelihood estimation or reconstruction error, but these methods can be computationally expensive, unreliable, and require retraining if the inlier data changes. These limitations hinder their ability to distinguish nominal from anomalous inputs efficiently, consistently, and robustly. We propose a reconstruction-free OOD detection method that leverages the forward diffusion trajectories of a Stein score-based denoising diffusion model (SBDDM). By capturing trajectory curvature via the estimated Stein score, our approach enables accurate anomaly scoring with only five diffusion steps. A single SBDDM pre-trained on a large, semantically aligned medical dataset generalizes effectively across multiple Near-OOD and Far-OOD benchmarks, achieving state-of-the-art performance while drastically reducing computational cost during inference. Compared to existing methods, SBDDM achieves a relative improvement of up to 10.43% and 18.10% for Near-OOD and Far-OOD detection, making it a practical building block for real-time, reliable computer-aided diagnosis.

Problem

• Detect out-of-distribution cases in medical imaging efficiently
• Reduce computational cost in anomaly detection without retraining
• Improve accuracy for Near-OOD and Far-OOD detection benchmarks

Innovation

• Uses Stein score-based denoising diffusion model
• Leverages forward diffusion trajectories curvature
• Requires only five diffusion steps