One-for-All Few-Shot Anomaly Detection via Instance-Induced Prompt Learning

Anomaly detection methods under the 'one-for-all' paradigm aim to develop a unified model capable of detecting anomalies across multiple classes. However, these approaches typically require a large number of normal samples for model training, which may not always be fulfilled in practice. Few-shot anomaly detection methods can address scenarios with limited data but require a tailored model for each class, following the 'one-for-one' paradigm. In this paper, we first proposed a one-for-all few-shot anomaly detection method with the assistance of vision-language models. Unlike previous CLIP-based methods that learn fixed prompts for each class, our method learns a class-shared prompt generator to adaptively generate suitable prompts for each instance. The prompt generator is trained by aligning the prompts with the visual space and utilizing guidance from general textual descriptions of normality and abnormality. In addition, we further propose a method to address the problem of how to retrieve valid similar features from the visual memory bank under the one-for-all paradigm. Extensive experimental results on MVTec and VisA demonstrate the superiority of our method in few-shot anomaly detection task under the one-forall paradigm. Our code is available in https://github.com/Vanssssry/ One-For-All-Few-Shot-Anomaly-Detection .