Industrial Surface Defect Detection via Diffusion Generation and Asymmetric Student-Teacher Network

2026-04-21

AI Summary

This work addresses the challenges of industrial surface defect detection, which are often hindered by scarce defect samples, long-tailed distributions, and the difficulty of localizing subtle anomalies against complex backgrounds. The authors propose an unsupervised method that innovatively integrates a denoising diffusion probabilistic model (DDPM) with an asymmetric teacher–student network. By training the DDPM exclusively on normal samples, the approach generates high-fidelity synthetic defect images along with their pixel-level annotations. Anomaly regions are further emphasized through a dual-stream teacher–student architecture, and the model is jointly optimized using Perlin noise masks, cosine similarity loss, and pixel-wise segmentation supervision. Evaluated on the MVTec AD dataset, the method achieves 98.4% image-level and 98.3% pixel-level AUROC, substantially outperforming existing unsupervised and mainstream deep learning approaches.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Unsupervised & Self-Supervised Learning
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Industrial surface defect detection often suffers from limited defect samples, severe long-tailed distributions, and difficulties in accurately localizing subtle defects under complex backgrounds. To address these challenges, this paper proposes an unsupervised defect detection method that integrates a Denoising Diffusion Probabilistic Model (DDPM) with an asymmetric teacher-student architecture. First, at the data level, the DDPM is trained solely on normal samples. By introducing constant-variance Gaussian perturbations and Perlin noise-based masks, high-fidelity and physically consistent defect samples along with pixel-level annotations are generated, effectively alleviating the data scarcity problem. Second, at the model level, an asymmetric dual-stream network is constructed. The teacher network provides stable representations of normal features, while the student network reconstructs normal patterns and amplifies discrepancies between normal and anomalous regions. Finally, a joint optimization strategy combining cosine similarity loss and pixel-wise segmentation supervision is adopted to achieve precise localization of subtle defects. Experimental results on the MVTecAD dataset show that the proposed method achieves 98.4\% image-level AUROC and 98.3\% pixel-level AUROC, significantly outperforming existing unsupervised and mainstream deep learning methods. The proposed approach does not require large amounts of real defect samples and enables accurate and robust industrial defect detection and localization. \keywords{Industrial defect detection \and diffusion models \and data generation \and teacher-student architecture \and pixel-level localization}

Problem

• Industrial defect detection
• limited defect samples
• long-tailed distribution
• subtle defect localization
• complex backgrounds

Innovation

• diffusion models
• asymmetric teacher-student architecture
• synthetic defect generation
• pixel-level localization
• unsupervised anomaly detection