计算机科学 ›› 2023, Vol. 50 ›› Issue (12): 212-220.doi: 10.11896/jsjkx.221000183
张邹铨1, 张辉2, 吴天月1, 陈天才3
ZHANG Zouquan1, ZHANG Hui2, WU Tianyue1, CHEN Tiancai3
摘要: 工业产品表面异常检测是生产制造中不可或缺的环节。在实际工业生产中,普遍存在异常样本所占比例低且未知异常复杂多变等现象,进而造成在小样本数据集上过拟合、泛化能力不佳等一系列负面影响。近年来,标准化流思想为基于深度学习的工业图像异常检测带来了新途径,但标准化流的固有架构易导致模型表达能力不足。针对上述难点,提出了一种面向工业图像异常检测的连续密集标准化流模型。首先,设计一种基于对比学习的特征提取网络预训练策略,将模拟异常数据和少量真实异常数据加入对比学习任务中,并训练特征骨干网络AlexNet拉近或拉远特定样本间的距离;其次,设计连续密集标准化流模型,采用可逆变换的复合架构来构造密集流模块,增强生成式模型对分布的拟合能力。在MVTec AD和Magnetic Tile Defects以及自制的工业布匹数据集上的实验结果表明,与其他的异常检测模型相比,所提方法在3个数据集上的检测性能达到了最优或次优的水平。
中图分类号:
[1]CHALAPATHY R,CHAWLA S.Deep Learning for AnomalyDetection:A Survey[EB/OL].(2019-01-23)[2022-10-10].https://arxiv.org/pdf/1901.03407.pdf. [2]LV C K,SHEN F,ZHANG Z T,et al.Review of Image Ano-maly Detection[J].Acta Automatica Sinica,2022,48(6):1402-1428. [3]BERGMANN P,FAUSER M,SATTLEGGER D,et al.MVTec AD-A Comprehensive Real-World Dataset for Unsupervised Anomaly Detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Long Beach:IEEE Press,2019:9592-9600. [4]HUANG Y,QIU C,YUAN K.Surface Defect Saliency of Magnetic Tile[J].The Visual Computer,2020,36(1):85-96. [5]XU T,TIAN C Y,LIU C H.Deep Learning for AbnormalCrowd Behavior Detection:A Review[J].Computer Science,2021,48(9):125-134. [6]WANG Y S,YAO H X,SUN X S,et al.Representation Ability Research of Autoencoders in Deep Learning[J].Computer Science,2015,42(9):56-60,65. [7]BERGMANN P,LÖWE S,FAUSER M,et al.Improving Unsupervised Defect Segmentation by Applying Structural Similarity to Autoencoders[EB/OL].(2019-02-01)[2022-10-10].https://arxiv.org/pdf/1807.02011.pdf. [8]YU W Y,ZHANG Y,YAO H M,et al.Visual Inspection ofSurface Defects Based on Lightweight Reconstruction Network[J].Acta Automatica Sinica,2022,48(9):2175-2186. [9]ZAVRTANIK V,KRISTAN M,SKOČAJ D.DRAEM:A Discriminatively Trained Reconstruction Embedding for Surface Anomaly Detection[C]//Proceedings of the IEEE International Conference on Computer Vision.Montreal:IEEE Press,2021:8330-8339. [10]KIM D,JEONG D,KIM H,et al.Spatial Contrastive Learning for Anomaly Detection and Localization[J].IEEE Access,2022,10:17366-17376. [11]LI C L,SOHN K,YOON J,et al.Cutpaste:Self-SupervisedLearning for Anomaly Detection and Localization[C]//Procee-dings of the IEEE Conference on Computer Vision and Pattern Recognition.Nashville:IEEE Press,2021:9664-9674. [12]ULGER F,YUKSELS E,YILMAZ A.Anomaly Detection forSolder Joints using β-VAE[J].IEEE Transactions on Components,Packaging and Manufacturing Technology,2021,11(12):2214-2221. [13]SCHLEGL T,SEEBÖCK P,WALDSTEIN S M,et al.Unsupervised Anomaly Detection withGenerative Adversarial Networks to Guide Marker Discovery[C]//International Conference on Information Processing in Medical Imaging.Springer,2017:146-157. [14]XING H J,HAO Z.Novelty Detection Method Based on Global and Local Discriminative Adversarial Autoencoder[J].Compu-ter Science,2021,48(6):202-209. [15]KOBYZEV I,PRINCE S J D,BRUBAKER M A.NormalizingFlows:An Introduction and Review of Current Methods[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2020,43(11):3964-3979. [16]DINH L,KRUEGER D,BENGIO Y.NICE:Non-Linear Independent Components Estimation[EB/OL].(2015-04-10)[2022-10-10].https://arxiv.org/pdf/1410.8516.pdf. [17]HUANG C W,DINH L,COURVILLE A.Augmented Normalizing Flows:Bridging the Gap between Generative Flows and Latent Variable Models[EB/OL].(2020-02-17)[2022-10-10].https://arxiv.org/pdf/2002.07101.pdf. [18]RUDOLPH M,WANDT B,ROSENHAHN B.Same Same but Differnet:Semi-Supervised Defect Detection with Normalizing Flows[C]//Proceedings of the IEEE Winter Conference on Applications of Computer Vision.Waikoloa:IEEE Press,2021:1907-1916. [19]CIMPOI M,MAJI S,KOKKINOS I,et al.Describing Textures in the Wild[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Columbus:IEEE Press,2014:3606-3613. [20]KRIZHEVSKY A,SUTSKEVER I,HINTON G E.ImagenetClassification with Deep Convolutional Neural Networks[J].Communications of the ACM,2017,60(6):84-90. [21]KRUSE J,DETOMMASO G,KÖTHE U,et al.HINT:Hierarchical Invertible Neural Transport for Density Estimation and Bayesian Inference[C]//Proceedings of the 35th AAAI Confe-rence on Artificial Intelligence.2021:8191-8199. [22]DINH L,SOHL-DICKSTEIN J,BENGIO S.Density Estimation using Real NVP[EB/OL].(2017-02-27)[2022-07-08].https://arxiv.org/pdf/1605.08803.pdf. [23]DEFARD T,SETKOV A,LOESCH A,et al.Padim:A Patch Distribution Modeling Framework for Anomaly Detection and Localization[C]//International Conference on Pattern Recognition.Springer,2021:475-489. [24]AKCAY S,ATAPOUR-ABARGHOUEI A,BRECKON T P.Ganomaly:Semi-Supervised Anomaly Detection via Adversarial Training[C]//Asian Conference on Computer Vision.Springer,2018:622-637. [25]PANG G,DING C,SHEN C.Explainable Deep Few-Shot Ano-maly Detection with Deviation Networks[EB/OL].(2021-08-01)[2022-10-10].https://arxiv.org/pdf/2108.00462.pdf. [26]DING C,PANG G,SHEN C.Catching Both Gray and BlackSwans:Open-Set Supervised Anomaly Detection[C]//Procee-dings of the IEEE Conference on Computer Vision and Pattern Recognition.2022:7388-7398. [27]VENKATARAMANAN S,PENG K C,SINGH R V,et al.Attention Guided Anomaly Localization in Images[C]//European Conference on Computer Vision.2020:485-503. |
|