Semantic Iterative Reconstruction: One-Shot Universal Anomaly Detection

TL;DR

Semantic Iterative Reconstruction (SIR) introduces a universal anomaly detection framework that effectively detects medical anomalies across diverse domains using minimal normal samples and a single trained model.

Contribution

SIR is the first to enable a one-shot universal anomaly detection model that generalizes across multiple medical domains with minimal data.

Findings

SIR outperforms previous methods on nine medical benchmarks.

Achieves state-of-the-art results in one-shot and full-shot settings.

Operates effectively across multiple medical modalities without retraining.

Abstract

Unsupervised medical anomaly detection is severely limited by the scarcity of normal training samples. Existing methods typically train dedicated models for each dataset or disease, requiring hundreds of normal images per task and lacking cross-modality generalization. We propose Semantic Iterative Reconstruction (SIR), a framework that enables a single universal model to detect anomalies across diverse medical domains using extremely few normal samples. SIR leverages a pretrained teacher encoder to extract multi-scale deep features and employs a compact up-then-down decoder with multi-loop iterative refinement to enforce robust normality priors in deep feature space. The framework adopts a one-shot universal design: a single model is trained by mixing exactly one normal sample from each of nine heterogeneous datasets, enabling effective anomaly detection on all corresponding test sets without task-specific retraining. Extensive experiments on nine medical benchmarks demonstrate that SIR achieves state-of-the-art under all four settings -- one-shot universal, full-shot universal, one-shot specialized, and full-shot specialized -- consistently outperforming previous methods. SIR offers an efficient and scalable solution for multi-domain clinical anomaly detection.