MAGIC: Few-Shot Mask-Guided Anomaly Inpainting with Prompt Perturbation, Spatially Adaptive Guidance, and Context Awareness

TL;DR

MAGIC is a novel inpainting framework for few-shot anomaly generation that enhances diversity and realism using prompt perturbation, spatial guidance, and mask alignment, improving industrial quality control models.

Contribution

Introduces MAGIC, a fine-tuned inpainting method with three components to generate diverse, high-fidelity anomalies for robust downstream industrial applications.

Findings

MAGIC outperforms existing methods on various anomaly datasets.

The approach produces high-fidelity, diverse anomalies adhering to masks.

MAGIC improves downstream model robustness in industrial quality control.

Abstract

Few-shot anomaly generation is a key challenge in industrial quality control. Although diffusion models are promising, existing methods struggle: global prompt-guided approaches corrupt normal regions, and existing inpainting-based methods often lack the in-distribution diversity essential for robust downstream models. We propose MAGIC, a fine-tuned inpainting framework that generates high-fidelity anomalies that strictly adhere to the mask while maximizing this diversity. MAGIC introduces three complementary components: (i) Gaussian prompt perturbation, which prevents model overfitting in the few-shot setting by learning and sampling from a smooth manifold of realistic anomalies, (ii) spatially adaptive guidance that applies distinct guidance strengths to the anomaly and background regions, and (iii) context-aware mask alignment to relocate masks for plausible placement within the host object. Under consistent identical evaluation protocol, MAGIC outperforms state-of-the-art methods on diverse anomaly datasets in downstream tasks.