FADE: Few-shot/zero-shot Anomaly Detection Engine using Large Vision-Language Model

TL;DR

FADE leverages large vision-language models, specifically CLIP, to enable effective zero-shot and few-shot anomaly detection in industrial images by adapting multi-scale embeddings and ensemble text prompts.

Contribution

This paper introduces FADE, a novel approach that adapts CLIP for industrial anomaly detection, improving language-guided segmentation and combining vision and language cues.

Findings

Outperforms state-of-the-art in anomaly segmentation

Achieves pixel-AUROC of 89.6% in zero-shot

Achieves pixel-AUROC of 95.4% in 1-normal-shot

Abstract

Automatic image anomaly detection is important for quality inspection in the manufacturing industry. The usual unsupervised anomaly detection approach is to train a model for each object class using a dataset of normal samples. However, a more realistic problem is zero-/few-shot anomaly detection where zero or only a few normal samples are available. This makes the training of object-specific models challenging. Recently, large foundation vision-language models have shown strong zero-shot performance in various downstream tasks. While these models have learned complex relationships between vision and language, they are not specifically designed for the tasks of anomaly detection. In this paper, we propose the Few-shot/zero-shot Anomaly Detection Engine (FADE) which leverages the vision-language CLIP model and adjusts it for the purpose of industrial anomaly detection. Specifically, we improve language-guided anomaly segmentation 1) by adapting CLIP to extract multi-scale image patch embeddings that are better aligned with language and 2) by automatically generating an ensemble of text prompts related to industrial anomaly detection. 3) We use additional vision-based guidance from the query and reference images to further improve both zero-shot and few-shot anomaly detection. On the MVTec-AD (and VisA) dataset, FADE outperforms other state-of-the-art methods in anomaly segmentation with pixel-AUROC of 89.6% (91.5%) in zero-shot and 95.4% (97.5%) in 1-normal-shot. Code is available at https://github.com/BMVC-FADE/BMVC-FADE.