Search papers, labs, and topics across Lattice.
This paper introduces a novel unsupervised industrial defect detection method using a Denoising Diffusion Probabilistic Model (DDPM) to generate synthetic defect samples and an asymmetric teacher-student network for anomaly localization. The DDPM, trained on normal samples with Gaussian perturbations and Perlin noise, generates high-fidelity defect data to address data scarcity. The asymmetric network, optimized with cosine similarity and pixel-wise segmentation loss, achieves state-of-the-art performance on the MVTecAD dataset, demonstrating effective defect detection and localization without real defect samples.
Synthesizing realistic defect data with diffusion models and Perlin noise can dramatically improve unsupervised anomaly detection, achieving near-perfect AUROC scores on industrial surfaces.
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}
