AREPAS: Anomaly Detection in Fine-Grained Anatomy with Reconstruction-Based Semantic Patch-Scoring

2025-09-16

AI Summary

Existing generative anomaly detection methods struggle to distinguish subtle physiological variations from genuine pathological anomalies in fine anatomical structures (e.g., lungs), limiting localization and segmentation accuracy. To address this, we propose a reconstruction-based semantic image translation framework coupled with local patch-wise similarity scoring: first, a generative image-to-image translation network reconstructs anomaly-free reference images; then, pixel-level anomaly detection is achieved by computing similarity scores between semantically aligned local patches of the reconstructed and original images. Our method innovatively integrates semantic-guided reconstruction with fine-grained local matching, significantly enhancing robustness to subtle anatomical variations. Evaluated on infection lesion segmentation in chest CT and stroke lesion segmentation in brain MRI, our approach achieves Dice score improvements of 1.9% and 4.4%, respectively, over the best-performing reconstruction-based baseline—demonstrating superior accuracy and strong cross-modal generalizability.

Technology Category

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

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 query analysis, representation and understanding
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

Early detection of newly emerging diseases, lesion severity assessment, differentiation of medical conditions and automated screening are examples for the wide applicability and importance of anomaly detection (AD) and unsupervised segmentation in medicine. Normal fine-grained tissue variability such as present in pulmonary anatomy is a major challenge for existing generative AD methods. Here, we propose a novel generative AD approach addressing this issue. It consists of an image-to-image translation for anomaly-free reconstruction and a subsequent patch similarity scoring between observed and generated image-pairs for precise anomaly localization. We validate the new method on chest computed tomography (CT) scans for the detection and segmentation of infectious disease lesions. To assess generalizability, we evaluate the method on an ischemic stroke lesion segmentation task in T1-weighted brain MRI. Results show improved pixel-level anomaly segmentation in both chest CTs and brain MRIs, with relative DICE score improvements of +1.9% and +4.4%, respectively, compared to other state-of-the-art reconstruction-based methods.

Problem

• Detecting anomalies in fine-grained pulmonary anatomy
• Improving unsupervised lesion segmentation in medical images
• Addressing normal tissue variability in generative anomaly detection

Innovation

• Reconstruction-based semantic patch-scoring for anomalies
• Image-to-image translation for anomaly-free reconstruction
• Patch similarity scoring for precise anomaly localization