PyramidFlow: High-Resolution Defect Contrastive Localization using Pyramid Normalizing Flow

TL;DR

PyramidFlow is a novel high-resolution defect localization method using fully normalizing flows without pre-trained models, employing multi-scale fusion and contrastive localization to improve industrial defect detection accuracy.

Contribution

It introduces the first fully normalizing flow approach for high-resolution defect localization that does not rely on pre-trained models, enhancing visual performance and generalization.

Findings

Outperforms algorithms without external priors on MVTecAD

Achieves state-of-the-art results on BTAD scenarios

Effectively reduces intra-class variance in defect localization

Abstract

During industrial processing, unforeseen defects may arise in products due to uncontrollable factors. Although unsupervised methods have been successful in defect localization, the usual use of pre-trained models results in low-resolution outputs, which damages visual performance. To address this issue, we propose PyramidFlow, the first fully normalizing flow method without pre-trained models that enables high-resolution defect localization. Specifically, we propose a latent template-based defect contrastive localization paradigm to reduce intra-class variance, as the pre-trained models do. In addition, PyramidFlow utilizes pyramid-like normalizing flows for multi-scale fusing and volume normalization to help generalization. Our comprehensive studies on MVTecAD demonstrate the proposed method outperforms the comparable algorithms that do not use external priors, even achieving state-of-the-art performance in more challenging BTAD scenarios.