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

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 lowresolution 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.

[1]  Xinyi Gong,et al.  Deep Learning for Unsupervised Anomaly Localization in Industrial Images: A Survey , 2022, IEEE Transactions on Instrumentation and Measurement.

[2]  Nilesh A. Ahuja,et al.  Anomalib: A Deep Learning Library for Anomaly Detection , 2022, 2022 IEEE International Conference on Image Processing (ICIP).

[3]  Ye Zheng,et al.  FastFlow: Unsupervised Anomaly Detection and Localization via 2D Normalizing Flows , 2021, ArXiv.

[4]  Bodo Rosenhahn,et al.  Fully Convolutional Cross-Scale-Flows for Image-based Defect Detection , 2021, 2022 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV).

[5]  D. Skočaj,et al.  DRÆM – A discriminatively trained reconstruction embedding for surface anomaly detection , 2021, 2021 IEEE/CVF International Conference on Computer Vision (ICCV).

[6]  Kazuki Kozuka,et al.  CFLOW-AD: Real-Time Unsupervised Anomaly Detection with Localization via Conditional Normalizing Flows , 2021, 2022 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV).

[7]  B. Schölkopf,et al.  Towards Total Recall in Industrial Anomaly Detection , 2021, 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[8]  Gian Luca Foresti,et al.  VT-ADL: A Vision Transformer Network for Image Anomaly Detection and Localization , 2021, 2021 IEEE 30th International Symposium on Industrial Electronics (ISIE).

[9]  Danijel Skocaj,et al.  Mixed supervision for surface-defect detection: from weakly to fully supervised learning , 2021, Comput. Ind..

[10]  Romaric Audigier,et al.  PaDiM: a Patch Distribution Modeling Framework for Anomaly Detection and Localization , 2020, ICPR Workshops.

[11]  Bodo Rosenhahn,et al.  Same Same But DifferNet: Semi-Supervised Defect Detection with Normalizing Flows , 2020, 2021 IEEE Winter Conference on Applications of Computer Vision (WACV).

[12]  Sungroh Yoon,et al.  Patch SVDD: Patch-level SVDD for Anomaly Detection and Segmentation , 2020, ACCV.

[13]  Andrew Gordon Wilson,et al.  Why Normalizing Flows Fail to Detect Out-of-Distribution Data , 2020, NeurIPS.

[14]  Yedid Hoshen,et al.  Sub-Image Anomaly Detection with Deep Pyramid Correspondences , 2020, ArXiv.

[15]  Paul Bergmann,et al.  Uninformed Students: Student-Teacher Anomaly Detection With Discriminative Latent Embeddings , 2019, 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[16]  Carsten Steger,et al.  MVTec AD — A Comprehensive Real-World Dataset for Unsupervised Anomaly Detection , 2019, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[17]  Jure Skvarč,et al.  Segmentation-based deep-learning approach for surface-defect detection , 2019, Journal of Intelligent Manufacturing.

[18]  Takashi Matsubara,et al.  Deep Generative Model Using Unregularized Score for Anomaly Detection With Heterogeneous Complexity , 2018, IEEE Transactions on Cybernetics.

[19]  Prafulla Dhariwal,et al.  Glow: Generative Flow with Invertible 1x1 Convolutions , 2018, NeurIPS.

[20]  Carsten Steger,et al.  Improving Unsupervised Defect Segmentation by Applying Structural Similarity to Autoencoders , 2018, VISIGRAPP.

[21]  Kay Chen Tan,et al.  A Generic Deep-Learning-Based Approach for Automated Surface Inspection , 2018, IEEE Transactions on Cybernetics.

[22]  Luca Antiga,et al.  Automatic differentiation in PyTorch , 2017 .

[23]  Georg Langs,et al.  Unsupervised Anomaly Detection with Generative Adversarial Networks to Guide Marker Discovery , 2017, IPMI.

[24]  Samy Bengio,et al.  Density estimation using Real NVP , 2016, ICLR.

[25]  Thomas Brox,et al.  U-Net: Convolutional Networks for Biomedical Image Segmentation , 2015, MICCAI.

[26]  Yoshua Bengio,et al.  NICE: Non-linear Independent Components Estimation , 2014, ICLR.

[27]  Weibo Liu,et al.  A Small-Sized Object Detection Oriented Multi-Scale Feature Fusion Approach With Application to Defect Detection , 2022, IEEE Transactions on Instrumentation and Measurement.

[28]  Chengqun Wang,et al.  ES-Net: Efficient Scale-Aware Network for Tiny Defect Detection , 2022, IEEE Transactions on Instrumentation and Measurement.

[29]  Feng Li,et al.  DefectNet: Toward Fast and Effective Defect Detection , 2021, IEEE Transactions on Instrumentation and Measurement.

[30]  Yilin Wu,et al.  FPCB Surface Defect Detection: A Decoupled Two-Stage Object Detection Framework , 2021, IEEE Transactions on Instrumentation and Measurement.

[31]  Jiabin Zhang,et al.  CADN: A weakly supervised learning-based category-aware object detection network for surface defect detection , 2021, Pattern Recognit..

[32]  Alan Carle,et al.  Automatic differentiation , 2003 .