基于块金字塔记忆模块的无监督异常检测

doi:10.19678/j.issn.1000-3428.0066232

摘要/Abstract

摘要：

基于重建的无监督异常检测方法由于不需要异常样本和预训练模型，被广泛地应用到异常检测任务中。然而，在实际应用中由于卷积神经网络的泛化性，模型能够有效地重建异常，使得难以通过重建误差来检测异常。现有方法通过使用合适的记忆块存储正常数据，将异常特征转化为正常特征，从而抑制异常重建，但不同的异常区域差异较大，记忆块尺寸的选择不当会导致重建模糊和重建异常等问题。考虑到这类方法在重建模型中的优势，提出一种基于改进记忆块存储的无监督异常检测方法。通过增加块金字塔记忆模块来适应不同面积大小的异常，并且不同尺度的块记忆模块通过读取、聚合得到多特征图融合的输出特征图，能够最大限度地保留正常样本的特征信息，增强特征信息的存储与表达，从而更好地重建正常数据。同时，为了增强重建清晰度，减少重建异常，在重建网络中增加skip connection结构。最后引入SSIM损失函数，通过亮度、对比度和结构3个维度来增强图像重建效果，并作为异常判定指标的组成部分，提高异常检测的精度。实验结果表明，相较于原始基于块存储和读取的重建模型，该方法平均AUC高出1.5%，具有更优的检测效果。

关键词: 无监督学习, 异常检测, 记忆模块, 自编码器, 重建模型

Abstract:

Unsupervised anomaly detection based on reconstruction is widely used because it does not require additional anomaly samples and pretrained models during training. However, in practical applications, the model can effectively reconstruct anomalies because of the generalization of convolutional neural networks, making it difficult to detect anomalies based on reconstruction errors. Existing methods suppress abnormal reconstruction by using appropriate memory blocks to store normal data and transform abnormal features into normal features. However, different abnormal areas can vary significantly, and improper selection of the memory block size may result in problems of fuzzy and abnormal reconstruction. Considering the advantages of these methods in the reconstruction model, an unsupervised anomaly detection method based on improved memory block storage is proposed. This involved adding a block pyramid memory module to adapt to anomalies of different area sizes. Block memory modules of different scales are employed to obtain multiple feature maps through reading and aggregation and subsequently obtained the output feature map after fusion. This approach can retain the feature information of normal samples to the maximum extent and enhance the storage and expression of feature information to better reconstruct normal data. Simultaneously, to enhance the clarity of the reconstruction and reduce reconstruction anomalies, a skip connection structure is added to the reconstruction network.Finally, the SSIM loss function is introduced to enhance the image reconstruction effect through the three dimensions of brightness, contrast, and structure. It also served as a component of the anomaly determination index, improving the accuracy of anomaly detection. The experimental results indicate that this method achieves an average AUC that is 1.5% higher than that of the original reconstructed model based on memory block storage and reading and has a better detection effect.

Key words: unsupervised learning, anomaly detection, memory module, autoencoder, reconstruction model

鄢宁, 李岳阳, 罗海驰. 基于块金字塔记忆模块的无监督异常检测[J]. 计算机工程, 2023, 49(12): 304-310.

Ning YAN, Yueyang LI, Haichi LUO. Unsupervised Anomaly Detection Based on Block Pyramid Memory Module[J]. Computer Engineering, 2023, 49(12): 304-310.

http://www.ecice06.com/CN/Y2023/V49/I12/304

图/表 6

图1 记忆模块网络结构

Fig.1 Memory module network structure

图2 整体网络结构

Fig.2 Overall network structure

图3 块金字塔记忆模块网络结构

Fig.3 Block pyramid memory module network structure

图4 不同算法的检测结果对比

Fig.4 Comparison of detection results of different algorithms

参考文献 26

1	ZIMEK A, SCHUBERT E, KRIEGEL H P. A survey on unsupervised outlier detection in high-dimensional numerical data. Statistical Analysis and Data Mining, 2012, 5 (5): 363- 387. doi: 10.1002/sam.11161
2	ZHAI S F, CHENG Y, LU W N, et al. Deep structured energy based models for anomaly detection[EB/OL]. [2022-10-10]. https://arxiv.org/abs/1605.07717.pdf.
3	BO Z, QI S, MIN M R, et al. Deep autoencoding gaussian mixture model for unsupervised anomaly detection[C]//Proceedings of International Conference on Learning Representations. Washington D. C., USA: IEEE Press, 2018: 357-368.
4	BERGMANN P, FAUSER M, SATTLEGGER D, et al. Uninformed students: student-teacher anomaly detection with discriminative latent embeddings[EB/OL]. [2022-10-10]. https://arxiv.org/abs/1911.02357.pdf.
5	BERGMANN P, LÖWE S, FAUSER M, et al. Improving unsupervised defect segmentation by applying structural similarity to autoencoders[EB/OL]. [2022-10-10]. https://arxiv.org/abs/1807.02011.pdf.
6	TANG T W, KUO W H, LAN J H, et al. Anomaly detection neural network with dual auto-encoders GAN and its industrial inspection applications. Sensors, 2020, 20 (12): 3336.
7	COLLIN A S, DE VLEESCHOUWER C. Improved anomaly detection by training an autoencoder with skip connections on images corrupted with stain-shaped noise[C]//Proceedings of the 25th International Conference on Pattern Recognition. Washington D. C., USA: IEEE Press, 2021: 7915-7922.
8	AKCAY S, ATAPOUR-ABARGHOUEI A, BRECKON T P. GANomaly: semi-supervised anomaly detection via adversarial training[C]//Proceedings of ACCVʼ18. Berlin, Germany: Springer, 2019: 622-637.
9	AKÇAY S, ATAPOUR-ABARGHOUEI A, BRECKON T P. Skip-GANomaly: skip connected and adversarially trained encoder-decoder anomaly detection[C]//Proceedings of International Joint Conference on Neural Networks. Washington D. C., USA: IEEE Press, 2019: 1-8.
10	冯志涛. 基于生成对抗网络的表面缺陷异常检测[D]. 大连: 大连理工大学, 2021.
	FENG Z T. Anomaly detection of surface defects based on generative countermeasure network[D]. Dalian: Dalian University of Technology, 2021. (in Chinese)
11	HASAN M, CHOI J, NEUMANN J, et al. Learning temporal regularity in video sequences[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2016: 733-742.
12	ZHAO Y R, DENG B, SHEN C, et al. Spatio-temporal AutoEncoder for video anomaly detection[C]//Proceedings of the 25th ACM International Conference on Multimedia. New York, USA: ACM Press, 2017: 1933-1941.
13	GONG D, LIU L Q, LE V, et al. Memorizing normality to detect anomaly: memory-augmented deep autoencoder for unsupervised anomaly detection[C]//Proceedings of IEEE/CVF International Conference on Computer Vision. Washington D. C., USA: IEEE Press, 2020: 1705-1714.
14	PARK H, NOH J, HAM B. Learning memory-guided normality for anomaly detection[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2020: 14360-14369.
15	李建伏, 王思博, 宋国平. 基于出行偏好和路径长度的路径规划方法. 吉林大学学报(理学版), 2021, 59 (1): 107- 114. URL
	LI J F, WANG S B, SONG G P. Path planning method based on routing preference and path length. Journal of Jilin University(Science Edition), 2021, 59 (1): 107- 114. URL
16	HOU J L, ZHANG Y Y, ZHONG Q Y, et al. Divide-and-assemble: learning block-wise memory for unsupervised anomaly detection[C]//Proceedings of IEEE/CVF International Conference on Computer Vision. Washington D. C., USA: IEEE Press, 2022: 8771-8780.
17	RONNEBERGER O, FISCHER P, BROX T. U-net: convolutional networks for biomedical image segmentation. Berlin, Germany: Springer, 2015.
18	WANG Z, BOVIK A C, SHEIKH H R, et al. Image quality assessment: from error visibility to structural similarity. IEEE Transactions on Image Processing, 2004, 13 (4): 600- 612.
19	KINGMA D P, WELLING M. Auto-encoding variational Bayes[EB/OL]. [2022-10-10]. https://arxiv.org/abs/1312.6114.pdf.
20	SAKURADA M, YAIRI T. Anomaly detection using autoencoders with nonlinear dimensionality reduction[C]//Proceedings of the 2nd Workshop on Machine Learning for Sensory Data Analysis. New York, USA: ACM Press, 2014: 153-160.
21	SCHÖLKOPF B, PLATT J, HOFMANN T. Greedy layer-wise training of deep networks[C]//Proceedings of NIPSʼ07. Cambridge, USA: MIT Press, 2007: 153-160.
22	SABOKROU M, KHALOOEI M, FATHY M, et al. Adversarially learned one-class classifier for novelty detection[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2018: 3379-3388.
23	RADFORD A, METZ L, CHINTALA S. Unsupervised representation learning with deep convolutional generative adversarial networks[EB/OL]. [2022-10-10]. https://arxiv.org/abs/1511.06434.
24	GOODFELLOW I J, POUGET-ABADIE J, MIRZA M, et al. Generative adversarial networks[EB/OL]. [2022-10-10]. https://arxiv.org/abs/1406.2661.pdf.
25	BERGMANN P, FAUSER M, SATTLEGGER D, et al. MVTec AD: a comprehensive real-world dataset for unsupervised anomaly detection[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2020: 9584-9592.
26	KINGMA D P, BA J. Adam: a method for stochastic optimization[EB/OL]. [2022-10-10]. https://arxiv.org/abs/1412.6980.pdf.

[1]	朱孟栩, 张文豪, 李国洪, 顾行发, 余涛, 郑逢杰, 张丽丽, 吴俣, 邴芳飞, 唐健雄. 基于卷积神经网络的高分六号卫星多光谱图像压缩[J]. 计算机工程, 2023, 49(9): 287-294.
[2]	王书朋, 何引弟. 融合特征注意力机制的非均匀光照图像增强算法[J]. 计算机工程, 2023, 49(8): 232-239.
[3]	陈仲磊, 伊鹏, 陈祥, 胡涛. 基于集成学习的系统调用实时异常检测框架[J]. 计算机工程, 2023, 49(6): 162-169,179.
[4]	王爱玲, 马文臻, 邹自明, 钟佳. 基于领域自适应的卫星工程参数异常检测[J]. 计算机工程, 2023, 49(5): 29-37,47.
[5]	张子宣, 宗学军, 何戡, 连莲. 基于CVAE-CatBoost的工业控制网络异常流量检测研究[J]. 计算机工程, 2023, 49(5): 173-180.
[6]	王禹博, 陈利锋, 许卫霞. 结合多解码器与两阶段通道选择的异常检测方法[J]. 计算机工程, 2023, 49(3): 37-48.
[7]	袁立宁, 胡皓, 刘钊. 基于多通道图卷积自编码器的图表示学习[J]. 计算机工程, 2023, 49(2): 150-160,174.
[8]	刘强, 张颖, 周卫祥, 蒋先涛, 周薇娜, 周谋国. 自适应类增量学习的物联网入侵检测系统[J]. 计算机工程, 2023, 49(2): 169-174.
[9]	何悦, 陈广胜, 景维鹏, 徐泽堃. 基于深度多相似性哈希方法的遥感图像检索[J]. 计算机工程, 2023, 49(2): 206-212.
[10]	富坤, 孙明磊, 郝玉涵, 刘赢华. 基于对抗训练的伪标签约束自编码器[J]. 计算机工程, 2023, 49(11): 123-130.
[11]	董卫宇, 李海涛, 王瑞敏, 任化娟, 孙雪凯. 基于堆叠卷积注意力的网络流量异常检测模型[J]. 计算机工程, 2022, 48(9): 12-19.
[12]	闫静, 张雪英, 李凤莲, 陈桂军, 黄丽霞. 结合栈式监督AE与可变加权ELM的回归预测模型[J]. 计算机工程, 2022, 48(8): 62-69,76.
[13]	江雨燕, 邵金, 李平. 融合自动权重学习的深度子空间聚类[J]. 计算机工程, 2022, 48(8): 77-84,97.
[14]	孙嘉, 张建辉, 卜佑军, 陈博, 胡楠, 王方玉. 基于CNN-BiLSTM模型的日志异常检测方法[J]. 计算机工程, 2022, 48(7): 151-158,167.
[15]	生龙, 袁丽娜, 武南南, 姬少培. 基于GSA与DE优化混合核ELM的网络异常检测模型[J]. 计算机工程, 2022, 48(6): 146-153.

选择文件类型/文献管理软件名称

选择包含的内容