Prototypical Learning Guided Context-Aware Segmentation Network for Few-Shot Anomaly Detection

2025-12-17

AI Summary

To address domain shift between pre-trained features and target scenarios in few-shot anomaly detection (FSAD), this paper proposes Prototype-guided Context-aware Segmentation Network (PCSNet). PCSNet introduces a novel prototype-based feature adaptation mechanism, integrated with a context-aware segmentation architecture enhanced by pseudo-anomaly synthesis and a pixel-level discrepancy classification loss, enabling fine-grained, pixel-level anomaly localization using only a few normal samples. Its core innovations lie in prototype-driven feature alignment to mitigate domain shift and contextual modeling to enhance local discriminability. Under the 8-shot setting on MVTec and MPDD benchmarks, PCSNet achieves image-level AUROC scores of 94.9% and 80.2%, respectively. Furthermore, it demonstrates strong generalization and practical utility in an industrial quality inspection task for automotive plastic components.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Security and Privacy: Large-scale security measurements

Abstract

Few-shot anomaly detection (FSAD) denotes the identification of anomalies within a target category with a limited number of normal samples. Existing FSAD methods largely rely on pre-trained feature representations to detect anomalies, but the inherent domain gap between pre-trained representations and target FSAD scenarios is often overlooked. This study proposes a Prototypical Learning Guided Context-Aware Segmentation Network (PCSNet) to address the domain gap, thereby improving feature descriptiveness in target scenarios and enhancing FSAD performance. In particular, PCSNet comprises a prototypical feature adaption (PFA) sub-network and a context-aware segmentation (CAS) sub-network. PFA extracts prototypical features as guidance to ensure better feature compactness for normal data while distinct separation from anomalies. A pixel-level disparity classification loss is also designed to make subtle anomalies more distinguishable. Then a CAS sub-network is introduced for pixel-level anomaly localization, where pseudo anomalies are exploited to facilitate the training process. Experimental results on MVTec and MPDD demonstrate the superior FSAD performance of PCSNet, with 94.9% and 80.2% image-level AUROC in an 8-shot scenario, respectively. Real-world applications on automotive plastic part inspection further demonstrate that PCSNet can achieve promising results with limited training samples. Code is available at https://github.com/yuxin-jiang/PCSNet.

Problem

• Addresses domain gap in few-shot anomaly detection
• Improves feature descriptiveness for target scenarios
• Enhances anomaly localization with limited normal samples

Innovation

• Prototypical feature adaption sub-network for domain gap reduction
• Pixel-level disparity classification loss for anomaly distinction
• Context-aware segmentation sub-network for precise anomaly localization