PatchFlow: Leveraging a Flow-Based Model with Patch Features

2026-02-05

AI Summary

This work addresses the performance degradation in surface defect detection for die-cast components caused by the mismatch between generic pre-trained features and the distinct distribution of industrial images. To this end, we propose an unsupervised zero-shot anomaly detection method that uniquely integrates local neighborhood-aware patch features with a normalizing flow model, complemented by a lightweight adapter module to effectively bridge the domain gap between pre-trained representations and industrial scenarios. Extensive experiments on the MVTec AD, VisA, and a private die-casting dataset demonstrate the superiority of our approach, achieving image-level AUROC scores of 99.28%, 96.48%, and 95.77%, respectively, and reducing error rates by 20%–28.2% compared to existing methods, thereby confirming its enhanced accuracy and generalization capability.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Calibration & Uncertainty Quantification
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Die casting plays a crucial role across various industries due to its ability to craft intricate shapes with high precision and smooth surfaces. However, surface defects remain a major issue that impedes die casting quality control. Recently, computer vision techniques have been explored to automate and improve defect detection. In this work, we combine local neighbor-aware patch features with a normalizing flow model and bridge the gap between the generic pretrained feature extractor and industrial product images by introducing an adapter module to increase the efficiency and accuracy of automated anomaly detection. Compared to state-of-the-art methods, our approach reduces the error rate by 20\% on the MVTec AD dataset, achieving an image-level AUROC of 99.28\%. Our approach has also enhanced performance on the VisA dataset , achieving an image-level AUROC of 96.48\%. Compared to the state-of-the-art models, this represents a 28.2\% reduction in error. Additionally, experiments on a proprietary die casting dataset yield an accuracy of 95.77\% for anomaly detection, without requiring any anomalous samples for training. Our method illustrates the potential of leveraging computer vision and deep learning techniques to advance inspection capabilities for the die casting industry

Problem

• die casting
• surface defects
• anomaly detection
• computer vision
• industrial inspection

Innovation

• normalizing flow
• patch features
• adapter module
• anomaly detection
• die casting