OSAGEN: Object-Aware Mask Priors and Multistage Decoupled Diffusion for Industrial Anomaly Generation

2026-07-31

AI Summary

Industrial anomaly detection and localization are hindered by the scarcity of real anomalous samples and pixel-level annotations. To address this challenge, this work proposes a three-stage decoupled diffusion framework that sequentially models normal appearance, coarse-grained defect representations, and fine-grained mask calibration. The method introduces an object-aware mask prior to synthesize high-quality image-mask pairs of anomalies. Innovatively, it integrates Query-Based Guidance (QBG) to inject object structural information and employs an Instance-Specific Constraint (ISC) to regulate the sampling process, thereby enhancing structural consistency and suppressing spurious anomaly diffusion. Evaluated on the MVTec AD and VisA datasets, the approach achieves AP-P/F1-P scores of 88.1/82.2 and 68.5/66.1, respectively, significantly outperforming existing methods.

Abstract

Industrial anomaly detection and localization are limited by scarce real anomalies and pixel-level annotations, a bottleneck that synthetic image-mask pairs can alleviate. However, existing few-shot mask-guided generation may over-follow mask geometry, produce weak anomalies, or use condition masks incompatible with the current object instance. We propose OSAGEN, which combines object-aware mask priors with multistage decoupled diffusion. Its three-stage adaptation sequentially learns normal appearance, defect appearance under coarse conditions, and fine-grained mask calibration, improving defect realization and local control. QBG injects object structure from a matched normal image into mask diffusion to produce object-aware priors, while ISC restricts anomaly propagation and preserves normal content during sampling. A lightweight materialization step recovers pixel-level labels aligned with the realized defects. On MVTec AD and VisA, OSAGEN achieves AP-P/F1-P scores of 88.1/82.2 and 68.5/66.1, respectively, under a unified downstream localization protocol. The code will be released upon acceptance.

Problem

• industrial anomaly detection
• anomaly generation
• mask-guided synthesis
• few-shot learning
• pixel-level annotation

Innovation

• object-aware mask priors
• multistage decoupled diffusion
• anomaly generation
• industrial anomaly detection
• mask-guided synthesis