Overview of AIGC Models for Industry Defect Detection

As a pivotal component of smart manufacturing, industrial defect detection has long faced core challenges such as the scarcity of defect samples and limited model generalization. AIGC (Artificial Intelligence Generated Content) models, leveraging their deep feature learning and data generation capabilities, provide a novel technical pathway to transcend conventional detection paradigms. This paper systematically reviews research progress and application scenarios of AIGC methodologies based on Generative Adversarial Networks (GAN), Transformer architectures, and Diffusion models in industrial defect detection. By analyzing technical frameworks encompassing high-fidelity defect generation, few-shot learning, multimodal fusion, and unsupervised anomaly detection, we reveal the unique advantages of AIGC models in enhancing detection accuracy and improving model generalization. This review aims to provide systematic references for theoretical innovation and practical implementation of AIGC technology in industrial quality inspection.