Unsupervised Anomaly Detection with an Enhanced Teacher for Student-Teacher Feature Pyramid Matching

2025-12-20

AI Summary

To address the insufficient multi-scale feature modeling and localization consistency in unsupervised anomaly detection, this paper proposes an enhanced teacher–student feature pyramid matching framework. We fine-tune a pre-trained ResNet-18 on MVTec-AD to construct a high-capacity teacher network and, for the first time, design it to guide a student network in aligning hierarchical feature pyramids—enabling joint optimization of image-level and pixel-level anomaly responses. The method integrates transfer learning, knowledge distillation, cross-level feature alignment, and unsupervised reconstruction error modeling. Evaluated on MVTec-AD, it achieves state-of-the-art performance: 0.971 image-level AUROC and 0.977 pixel-level AUROC—significantly outperforming existing unsupervised approaches. These results validate the effectiveness and generalizability of the proposed multi-scale teacher–student alignment mechanism.

Technology Category

• Machine Learning: Unsupervised & Self-Supervised Learning
• Computer Vision: Adversarial Attacks & Robustness
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Anomaly detection or outlier is one of the challenging subjects in unsupervised learning . This paper is introduced a student-teacher framework for anomaly detection that its teacher network is enhanced for achieving high-performance metrics . For this purpose , we first pre-train the ResNet-18 network on the ImageNet and then fine-tune it on the MVTech-AD dataset . Experiment results on the image-level and pixel-level demonstrate that this idea has achieved better metrics than the previous methods . Our model , Enhanced Teacher for Student-Teacher Feature Pyramid (ET-STPM), achieved 0.971 mean accuracy on the image-level and 0.977 mean accuracy on the pixel-level for anomaly detection.

Problem

• Develops unsupervised anomaly detection using student-teacher framework
• Enhances teacher network for improved feature pyramid matching
• Achieves high accuracy in image and pixel-level anomaly detection

Innovation

• Enhanced teacher network in student-teacher framework
• Pre-trained ResNet-18 fine-tuned on MVTech-AD dataset
• Feature pyramid matching for anomaly detection metrics