基于条件轻量级神经网络的视频入侵检测算法

doi:10.19678/j.issn.1000-3428.0066529

摘要/Abstract

摘要：

机场安防场景需要快速准确地对入侵目标进行检测及报警，现有的算法计算量大，无法满足实时处理需求。针对该问题，结合传统高斯混合模型的前景提取算法与轻量级神经网络，设计基于条件计算的轻量级入侵检测算法。在前景提取阶段使用基于信息熵的自适应学习因子更新算法动态更新高斯混合模型的学习因子，减小高斯混合模型面对镜头突入时造成模型失效的概率。目标检测阶段以ResNeXt作为主干网络，将小型ResNet作为策略网络，使混合感受野的深度可分离卷积作为残差块，设计基于条件计算的轻量级神经网络，降低网络推理时的计算量。实验结果表明，该算法在监控视频数据集和OTB100数据集上的误检率分别4.4%、9.2%，漏检率分别为2.3%、9.8%，与Faster-YOLO等传统目标检测算法相比，该算法在保证检测精度的情况下使检测速度平均提高了2.6倍。

关键词: 视频入侵检测, 高斯混合模型, 信息熵, 轻量级神经网络, 条件计算, 模型剪裁

Abstract:

Airport security needs to quickly and accurately detect and alarm an intrusion target. Existing algorithms require a large number of calculations and cannot meet real-time processing requirements. To address this problem, a lightweight intrusion detection algorithm based on conditional computation is designed by combining the traditional Gaussian mixture model foreground extraction algorithm with a lightweight neural network. In the foreground extraction stage, an adaptive learning factor update algorithm based on information entropy is used to dynamically update the learning factors of the Gaussian mixture model, thereby reducing the probability of model failure caused by the Gaussian mixture model in case of lens break-ins. In the target detection stage, ResNeXt is used as the backbone network; a small ResNet is used as the policy network; and a depth-wise separable convolution of the mixed receptive field is used as the residual block. A lightweight neural network based on conditional calculations is designed to reduce the number of calculations required for network inference. In the experimental results, the false detection rates of the algorithm on the surveillance video and OTB100 datasets were 4.4% and 9.2%, respectively, and the missed detection rates were 2.3% and 9.8%, respectively. Compared with traditional target detection methods, the proposed algorithm increases detection speed by an average of 2.6 times while ensuring detection accuracy.

Key words: video intrusion detection, Gaussian mixture model, information entropy, lightweight neural network, conditional computation, model tailoring

陈天宇, 楚程钱, 万思远, 万永菁, 孙静. 基于条件轻量级神经网络的视频入侵检测算法[J]. 计算机工程, 2023, 49(12): 152-160.

Tianyu CHEN, Chengqian CHU, Siyuan WAN, Yongjing WAN, Jing SUN. Video Intrusion Detection Algorithm Based on Conditional Lightweight Neural Network[J]. Computer Engineering, 2023, 49(12): 152-160.

https://www.ecice06.com/CN/Y2023/V49/I12/152

图/表 13

图1 入侵检测算法流程

Fig.1 Procedure of intrusion detection algorithm

图2 前景运动目标提取流程示意图

Fig.2 Schematic diagram of the procedure of foreground moving target extraction

图3 残差模块示意图

Fig.3 Schematic diagram of the residual block

图4 轻量级神经网络示意图

Fig.4 Schematic diagram of lightweight neural network

图5 信息熵帧差值分布统计图

Fig.5 Statistical diagram of information entropy frame difference distribution

参考文献 27

1	王波. 机场飞行区智能周界安防系统设计探讨. 科学技术创新, 2019, 78 (29): 104- 105. URL
	WANG B. Discussion on the design of intelligent perimeter security system in airport flight area. Scientific and Technological InnovationInformation, 2019, 78 (29): 104- 105. URL
2	段其昌, 赵钦波, 杨源飞. 一种基于特征匹配的目标入侵检测方法. 计算机应用, 2012, 32 (S1): 126-127, 133. URL
	DUAN Q C, ZHAO Q B, YANG Y F. Trespass detection method based on feature matching. Journal of Computer Applications, 2012, 32 (S1): 126-127, 133. URL
3	ANSARI S A, ZAFAR A. A fusion of dolphin swarm optimization and improved sine cosine algorithm for automatic detection and classification of objects from surveillance videos. Measurement, 2022, 192, 110921. doi: 10.1016/j.measurement.2022.110921
4	杜文汉, 李东兴, 王倩楠, 等. 融合改进帧差和边缘提取算法的运动目标检测. 科学技术与工程, 2022, 22 (5): 1944- 1949. doi: 10.3969/j.issn.1671-1815.2022.05.027
	DU W H, LI D X, WANG Q N, et al. Moving target detection based on improved frame difference and edge extraction algorithm. Science Technology and Engineering, 2022, 22 (5): 1944- 1949. doi: 10.3969/j.issn.1671-1815.2022.05.027
5	STAUFFER C, GRIMSON W E L. Adaptive background mixture models for real-time tracking[C]//Proceedings of 1999 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2002: 246-252.
6	VAN DROOGENBROECK M, PAQUOT O. Background subtraction: experiments and improvements for ViBe[C]//Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops. Washington D. C., USA: IEEE Press, 2012: 32-37.
7	GIRSHICK R, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of 2014 IEEE Conference on Computer Vision and Pattern Recognition. New York, USA: ACM Press, 2014: 580-587.
8	REN S Q, HE K M, GIRSHICK R, et al. Faster R-CNN: towards real-time object detection with region proposal networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39 (6): 1137- 1149. doi: 10.1109/TPAMI.2016.2577031
9	HE K M, GKIOXARI G, DOLLÁR P, et al. Mask R-CNN[C]//Proceedings of IEEE International Conference on Computer Vision. Washington D. C., USA: IEEE Press, 2017: 2980-2988.
10	REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: unified, real-time object detection[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2016: 779-788.
11	REDMON J, FARHADI A. YOLOv3: an incremental improvement[EB/OL]. [2022-11-08]. https://arxiv.org/abs/1804.02767.
12	YIN Y, LI H, FU W. Faster-YOLO: an accurate and faster object detection method. Digital Signal Processing, 2020, 102, 102756. doi: 10.1016/j.dsp.2020.102756
13	LIU W, ANGUELOV D, ERHAN D, et al. SSD: single shot multibox detector[C]//Proceedings of European Conference on Computer Vision. Berlin, Germany: Springer, 2016: 21-37.
14	LIN T Y, GOYAL P, GIRSHICK R, et al. Focal loss for dense object detection[C]//Proceedings of IEEE International Conference on Computer Vision. Washington D. C., USA: IEEE Press, 2017: 2999-3007.
15	XU S, WANG X, LV W, et al. PP-YOLOE: an evolved version of YOLO[EB/OL]. [2022-11-08]. https://arxiv.org/abs/2203.16250.
16	WANG C Y, BOCHKOVSKIY A, LIAO H Y M. YOLOv7: trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[EB/OL]. [2022-11-08]. https://arxiv.org/abs/2207.02696.
17	徐增敏, 陈凯, 郭威伟, 等. 面向轻量级卷积网络的激活函数与压缩模型. 计算机工程, 2022, 48 (5): 242- 250. URL
	XU Z M, CHEN K, GUO W W, et al. Activation function and compression model for lightweight convolutional network. Computer Engineering, 2022, 48 (5): 242- 250. URL
18	STAUFFER C, GRIMSON W E L. Learning patterns of activity using real-time tracking. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000, 22 (8): 747- 757. doi: 10.1109/34.868677
19	XIE S N, GIRSHICK R, DOLLÁR P, et al. Aggregated residual transformations for deep neural networks[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2017: 5987-5995.
20	TAN M, LE Q V. MixConv: mixed depthwise convolutional kernels[EB/OL]. [2022-11-08]. https://arxiv.org/abs/1907.09595.
21	WU Z X, NAGARAJAN T, KUMAR A, et al. BlockDrop: dynamic inference paths in residual networks[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2018: 8817-8826.
22	HOWARD A G, ZHU M, CHEN B, et al. MobileNets: efficient convolutional neural networks for mobile vision applications[EB/OL]. [2022-11-08]. https://arxiv.org/abs/1704.04861.
23	陈涛, 陈天宇, 万永菁, 等. 一种改进的适用于监控视频的轻量级入侵检测算法及其应用. 华东理工大学学报(自然科学版), 2021, 47 (6): 734- 741. URL
	CHEN T, CHEN T Y, WAN Y J, et al. An improved lightweight intrusion detection algorithm for surveillance video and its application. Journal of East China University of Science and Technology (Natural Science Edition), 2021, 47 (6): 734- 741. URL
24	HAN K, WANG Y H, TIAN Q, et al. GhostNet: more features from cheap operations[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2020: 1577-1586.
25	DING X H, GUO Y C, DING G G, et al. ACNet: strengthening the kernel skeletons for powerful CNN via asymmetric convolution blocks[C]//Proceedings of IEEE/CVF International Conference on Computer Vision. Washington D. C., USA: IEEE Press, 2020: 1911-1920.
26	YANG B, BENDER G, LE Q V, et al. CondConv: conditionally parameterized convolutions for efficient inference[EB/OL]. [2022-11-08]. https://arxiv.org/abs/1904.04971.
27	LI D, HU J, WANG C H, et al. Involution: inverting the inherence of convolution for visual recognition[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2021: 12316-12325.

[1]	陈何雄, 罗宇薇, 韦云凯, 郭威, 杭菲璐, 何映军, 杨宁. 基于联邦学习的SDN异常流量协同检测技术[J]. 计算机工程, 2023, 49(3): 168-176.
[2]	孙福禄, 王宇嘉, 刘子怡. 基于节点引力与鱼记忆的社区检测算法[J]. 计算机工程, 2022, 48(5): 104-111.
[3]	王文欣, 贺煜航, 陈刚. 基于EM路由算法的医学图像分割UCaps网络[J]. 计算机工程, 2022, 48(2): 268-274.
[4]	胡高珍, 徐胜军, 孟月波, 刘光辉, 冯峰, 段中兴. 基于边缘约束局部区域MRF的图像分割方法[J]. 计算机工程, 2021, 47(6): 253-261,270.
[5]	朱映波, 赵阳洋, 王佩, 尹凯, 王振宇. 融合马尔科夫决策过程与信息熵的对话策略[J]. 计算机工程, 2021, 47(3): 284-290.
[6]	王旭, 陈永乐, 王庆生, 陈俊杰. 结合特征选择与集成学习的密码体制识别方案[J]. 计算机工程, 2021, 47(1): 139-145,153.
[7]	张墨华, 彭建华. 面向图像先验建模的可扩展高斯混合模型[J]. 计算机工程, 2020, 46(4): 220-227.
[8]	赵军, 朱荽, 杨雯璟, 许彦辉, 庞宇. 一种基于密度峰值聚类的图像分割算法[J]. 计算机工程, 2020, 46(2): 274-278,285.
[9]	姚博凡, 邓红平, 蔡铭. 基于随机抽样GMM的城市交通运行状态模式分类[J]. 计算机工程, 2020, 46(12): 36-42.
[10]	郑文秀, 赵峻毅, 文心怡, 姚引娣. 基于瓶颈复合特征的声学模型建立方法[J]. 计算机工程, 2020, 46(11): 301-305,314.
[11]	邓路佳,刘平山. 基于GMM-FMs的广告点击率预测研究[J]. 计算机工程, 2019, 45(5): 122-126.
[12]	王万耀,段先华,徐丹,於跃成,黄炜亮. 基于显著性的Grabcut图像分割方法[J]. 计算机工程, 2018, 44(7): 230-236,243.
[13]	朱晓妤,严云洋,刘以安,高尚兵. 基于深度森林模型的火焰检测[J]. 计算机工程, 2018, 44(7): 264-270.
[14]	薛敏,方勇,黄诚,刘亮. 源代码中的API密钥自动识别方法[J]. 计算机工程, 2018, 44(6): 117-121,129.
[15]	黄玉洁,唐作其,梁静. 基于信息熵与三参数区间的信息安全风险评估[J]. 计算机工程, 2018, 44(12): 178-183.

选择文件类型/文献管理软件名称

选择包含的内容