A cyclic diffusion framework for structure-authentic and annotation-disentangled anomaly generation

doi:10.1016/j.neucom.2026.132839

A cyclic diffusion framework for structure-authentic and annotation-disentangled anomaly generation

2026-01-22

0

PRE

AI

DOI：10.1016/j.neucom.2026.132839

Original

Original request for help

Save

Abstract

En 中文

Visual anomaly inspection is critical in manufacturing, yet hampered by the scarcity of real anomaly samples for training robust detectors. Synthetic data generation presents a viable strategy for data augmentation; however, current methods remain constrained by two principal limitations: 1) the generation of anomalies that are structurally inconsistent with the normal background, and 2) the presence of undesirable feature entanglement between synthesized images and their corresponding annotation masks, which undermines the perceptual realism of the output. This paper introduces Cyclic Diffusion, CycDiff, a novel cross-domain generative framework designed to simultaneously synthesize high-fidelity anomaly images and their pixel-level annotation masks, explicitly addressing these challenges. CycDiff employs a unique architecture, cycling through distinct modules for feature separation, connection, and merging. Specifically, a domain-decoupled attention mechanism mitigates feature entanglement by enhancing image while annotation features independently, and a semantic score map alignment module ensures structural authenticity by coherently integrating anomaly foregrounds. CycDiff offers flexible control via text prompts and optional graphical guidance. Extensive experiments demonstrate that CycDiff significantly outperforms state-of-the-art methods in diversity and authenticity, leading to significant improvements in downstream anomaly detection performance.

A cyclic diffusion framework for structure-authentic and annotation-disentangled anomaly generation

Abstract

Journal

Organization