Exploring Large Vision-Language Models for Robust and Efficient Industrial Anomaly Detection
Proposes a novel approach (CLAD) leveraging large vision-language models with contrastive cross-modal training for improved industrial anomaly detection and localization, enhancing interpretability and robustness.