[1]戚小莎,曾 静,吉根林.双交叉注意力自编码器改进视频异常检测[J].南京师大学报(自然科学版),2023,46(01):110-119.[doi:10.3969/j.issn.1001-4616.2023.01.015]
　Qi Xiaosha,Zeng Jing,Ji Genlin.Improved Video Anomaly Detection with Dual Criss-Cross Attention Auto Encoder[J].Journal of Nanjing Normal University(Natural Science Edition),2023,46(01):110-119.[doi:10.3969/j.issn.1001-4616.2023.01.015]
点击复制

双交叉注意力自编码器改进视频异常检测()
分享到：

《南京师大学报（自然科学版）》[ISSN:1001-4616/CN:32-1239/N]

卷:
第46卷
期数:
2023年01期
页码:
110-119
栏目:
计算机科学与技术
出版日期:
2023-03-15

文章信息/Info

Title:
Improved Video Anomaly Detection with Dual Criss-Cross Attention Auto Encoder
文章编号:
1001-4616(2023)01-0110-10
作者:
戚小莎1; 曾 静2; 吉根林2
(1.南京师范大学数学科学学院，江苏 南京 210023)
(2.南京师范大学计算机与电子信息/人工智能学院，江苏 南京 210023)
Author(s):
Qi Xiaosha1; Zeng Jing2; Ji Genlin2
(1.School of Mathematical Sciences, Nanjing Normal University, Nanjing 210023, China)
(2.School of Computer and Electronic Information/Artificial Intelligence, Nanjing Normal University, Nanjing 210023, China)
关键词:
异常检测; 自编码器; 帧; 重构; 深度学习; 神经网络; 特征提取; 融合
Keywords:
anomaly detection;  auto encoder;  frame;  reconstruction;  deep learning;  neural network;  feature extraction;  fusion
分类号:
TP391
DOI:
10.3969/j.issn.1001-4616.2023.01.015
文献标志码:
A
摘要:
针对视频中包含的异常事件数量稀少,信息密集的特征容易被遗漏等问题,本文提出一种双交叉注意力自编码器的视频异常事件检测方法. 首先预处理视频集,提取视频帧中表观和运动特征,然后设计双交叉注意力模块融入自编码器中,使特征图在自编码器中能够更好地关联全局特征. 其次将提取后的特征放入各自的自编码器中学习正常行为,使含有正常事件的视频帧能被模型重构,含有异常事件的视频帧则无法被重构. 最后通过检测模型得到各个视频帧的重构误差从而进行异常事件判定. 该方法可以以局部特征关联全局特征的方式有效提高视频异常事件检测的准确率,通过在多个公开数据集中进行实验验证,证明该方法优于其他同类方法.
Abstract:
To solve the problems such as sparse quantity of abnormal events contained in video and information-intensive features are easy ommitted,this paper proposes a dual criss-cross attention auto encoder for video abnormal detection. Firstly,we preprocess the video to extract the apparent and motion features in the video,then design the dual criss-cross attention module and incorporate it into auto encoder,in this way,the features can better correlate the global features. Further,we put the extracted features into the respective auto encoders to learn normal behavior,in this way,video frames containing normal events can be reconstructed by the model and those containing abnormal events cannot be reconstructed. Finally,reconstruction errors of each video frame are obtained by the model to determine the abnormal events. This method can effectively improve the accuracy of abnormal event detection by correlating global features with local features,and it is proved to be better than other similar methods through experimental validation in several public datasets.

参考文献/References:

[1]ZAIGHAMZAHEER M,JIN H,LEE S,et al. A brief survey on contemporary methods for anomaly detection in videos[C]//Proceedings of International Conference on Information And Communication Technology Convergence. Allahabad,India,2019:472-473.
[2]CHIMAN D,DINESH K. A review of state-of-the-art techniques for abnormal human activity recognition[J]. Engineering applications of artificial intelligence,2019,77:21-45.
[3]SARAH A,GABRIEL P,SILVIO J,et al. Fight detection in video sequences based on multi-stream convolutional neural networks[C]//Proceedings of 32nd SIBGRAPI Conference on Graphics,Patterns and Images. Rio de Janeiro,Brazil,2019:8-15.
[4]ZHU C,WANG Y K,PU D B,et al. Multi-modality video representation for action recognition[J]. Journal on Big Data,2020,2(3):95-104.
[5]XU D M,HAN G G. Application of improved ViBe algorithm in vehicle detection[C]//International Conference on Artificial Intelligence and Pattern Recognition. Beijing,China,2021:199-204.
[6]ILG E,MAYER N,SAIKIA T,et al. FlowNet 2.0:Evolution of optical flow estimation with deep networks[C]//Proceedings of Conference on Computer Vision and Pattern Recognition. Hawaii,USA,2017:1647-1655.
[7]BUKALA A,KOZIARSKI M,CYGANEK B,et al. A study on pattern recognition with the histograms of oriented gradients in distorted and noisy images[J]. Journal of universal computer science,2020,26(4):454-478.
[8]SU Z B,LI W,MA Z,et al. An improved U-Net method for the semantic segmentation of remote sensing images[J]. Applied intelligence,2022,52:3276-3288.
[9]徐涛,田崇阳,刘才华. 基于深度学习的人群异常行为检测综述[J]. 计算机科学,2021,48(9):125-134.
[10]PILLAI G,SEN D. Anomaly detection in nonstationary videos using time-recursive differencing network-based prediction[J]. IEEE geoscience and remote sensing letters,2022,19:1-5.
[11]CHANG Y P,TU Z G,XIE W,et al. Video anomaly detection with spatio-temporal dissociation[J]. Pattern recognition,2022,122:2-13.
[12]LI N N,ZHONG J X,SHU X J,et al. Weakly-supervised anomaly detection in video surveillance via graph convolutional label noise cleaning[J]. Neurocomputing,2022,418:154-167.
[13]LUO W X,LIU W,GAO S H. Normal graph:spatial temporal graph convolutional networks based prediction network for skeleton based video anomaly detection[J]. Neurocomputing,2021,444:332-337.
[14]OUYANG Y Q,VICTOR S. Video anomaly detection by estimating likelihood of representations[C]//Proceedings of 25th International Conference on Pattern Recognition. Milan,Italy,2020:8984-8991.
[15]李自强,王正勇,陈洪刚,等. 基于外观和动作特征双预测模型的视频异常行为检测[J]. 计算机应用,2021,41(10):2997-3003.
[16]LI T,CHEN X Y,ZHU F S,et al. Two-stream deep spatial-temporal auto-encoder for surveillance video abnormal event detection[J]. Neurocomputing,2021,439:256-270.
[17]SUN C,JIA Y D,SONG H,et al. Adversarial 3D convolutional auto-encoder for abnormal event detection in videos[J]. IEEE transactions on multimedia,2021,23:3292-3305.
[18]ESQUIVEL E,ZAVALETA Z. An examination on autoencoder designs for anomaly detection in video surveillance[J]. IEEE access,2022,10:6208-6217.
[19]吕浩,易鹏飞,刘瑞,等. 用于视频异常检测的时序多尺度自编码器[J]. 图学学报,2022,43(2):223-229.
[20]CHAUDHARI S,MITHAL V,POLATKAN R,et al. An attentive survey of attention models[J]. ACM transactions on intelligent systems and technology,2021,12(53):1-32.
[21]SNEHASHIAS M,SRIJAN D,FRANCOIS B. DAM:dissimilarity attention module for weakly-supervised video anomaly detection[C]//Proceedings of IEEE International Conference on Advanced Video and Signal Based Surveillance. Washington,USA,2021:1-8.
[22]WANG C X,YAO Y X,YAO H. Video anomaly detection method based on future frame prediction and attention mechanism[C]//Proceedings of IEEE 11th Annual Computing and Communication Workshop and Conference. Online,2021:405-407.
[23]魏思倩,吉根林,许振,等. 利用注意力机制的多示例学习视频异常检测[J/OL]. 沈阳市,小型微型计算机系统,2021.(2021-10-18)[2022-04-05]. http://kns.cnki.net/kcms/detail/21.1106.tp.20211014.1237.002.html.
[24]FENG S T,ZHUO Z S,PAN D R,et al. CcNet:A cross-connected convolutional network for segmenting retinal vessels using multi-scale features[J]. Neurocomputing,2020,392:603-612.
[25]PIGA N,ONYSHCHUK Y,PASQUALE G,et al. ROFT:real-time optical flow-aided 6D object pose and velocity tracking[J]. IEEE robotics and automation letters,2022,7(1):159-166.
[26]CHEN W Y,PODSTRELENY P,CHENG W H,et al. Code generation from a graphical user interface via attention-based encoder-decoder model[J]. Multimedia systems,2022,28:121-130.
[27]ZHANG C X,HU Y H,ZHU X M. Anomaly detection for user behavior in wireless network based on cross entropy[C]//Proceedings of IEEE 12th International Conference on Ubiquitous Intelligence and Computing and IEEE 12th International Conference on Autonomic and Trusted Computing and IEEE 15th International Conference on Scalable Computing and Communications and Its Associated Workshops. Goyang,South Korea,2015:1258-1263.
[28]BERT D,DAVY N,LUC V G. Semantic instance segmentation with a discriminative loss function[EB/OL]. 2017. http://arxiv.org/pdf/1708.02551.
[29]LU C W,SHI J P,JIA J Y. Abnormal event detection at 150 FPS in MATLAB[C]//Proceedings of IEEE International Conference on Computer Vision. Portland,USA,2013:2720-2727.
[30]VIJAY M,LI W X,VIRAL B. Anomaly detection in crowded scenes[C]//Proceedings of Computer Vision & Pattern Recognition. San Francisco,USA,2010:1975-1981.
[31]李欣璐,吉根林,赵斌. 基于卷积自编码器分块学习的视频异常事件检测与定位[J]. 数据采集与处理,2021,36(3):489-497.
[32]YONG S,YONG H. Abnormal event detection in videos using spatiotemporal autoencoder[C]//Proceedings of the International Symposium in Neural Networks. Hokkaido,Japan,2017:189-196.
[33]LUO W X,LIU W,LIAN D Z,et al. Video anomaly detection with sparse coding inspired deep neural networks[J]. IEEE transactions on pattern analysis and machine intelligence,2021,43(3):1070-1084.
[34]RASHMIKE N,DAMMINDA A,DASWIN D,et al. Spatiotemporal anomaly detection using deep learning for real-time video surveillance[J]. IEEE transactions on industrial informatics,2020,16(1):393-402.
[35]XU Z,ZENG X Q,JI G L. Improved anomaly detection in surveillance videos with multiple probabilistic models inference[J]. Intelligent automation and soft computing,2022,31(3):1703-1717.

相似文献/References:

[1]李致远,朱求志,吴永焜,等.基于小波分析的无线传感网实时异常检测算法[J].南京师大学报(自然科学版),2014,37(01):87.
　Li Zhiyuan,Zhu Qiuzhi,Wu Yongkun,et al.Wavelet Analysis-Based Real-Time Anomaly Detection Algorithm for Wireless Sensor Network[J].Journal of Nanjing Normal University(Natural Science Edition),2014,37(01):87.

备注/Memo

备注/Memo:
收稿日期:2022-09-03.
基金项目:国家自然科学基金项目(41971343).
通讯作者:吉根林,博士,教授,研究方向:大数据分析与挖掘技术. E-mail:glji@njnu.edu.cn
更新日期/Last Update: 2023-03-15