CLIP-FSAC++: Few-Shot Anomaly Classification with Anomaly Descriptor Based on CLIP

TL;DR

CLIP-FSAC++ introduces a novel few-shot anomaly classification framework that enhances CLIP's representations through a cross-modality interaction module, improving anomaly detection accuracy in industrial scenarios with limited data.

Contribution

The paper proposes CLIP-FSAC++, a one-stage training framework with an anomaly descriptor that improves CLIP's adaptation for few-shot industrial anomaly classification.

Findings

Effective in 1, 2, 4, 8-shot settings on VisA and MVTEC-AD datasets.

Enhanced correlation between visual and text embeddings improves classification.

Outperforms existing few-shot anomaly detection methods.

Abstract

Industrial anomaly classification (AC) is an indispensable task in industrial manufacturing, which guarantees quality and safety of various product. To address the scarcity of data in industrial scenarios, lots of few-shot anomaly detection methods emerge recently. In this paper, we propose an effective few-shot anomaly classification (FSAC) framework with one-stage training, dubbed CLIP-FSAC++. Specifically, we introduce a cross-modality interaction module named Anomaly Descriptor following image and text encoders, which enhances the correlation of visual and text embeddings and adapts the representations of CLIP from pre-trained data to target data. In anomaly descriptor, image-to-text cross-attention module is used to obtain image-specific text embeddings and text-to-image cross-attention module is used to obtain text-specific visual embeddings. Then these modality-specific embeddings are used to enhance original representations of CLIP for better matching ability. Comprehensive experiment results are provided for evaluating our method in few-normal shot anomaly classification on VisA and MVTEC-AD for 1, 2, 4 and 8-shot settings. The source codes are at https://github.com/Jay-zzcoder/clip-fsac-pp