激光与光电子学进展, 2019, 56 (13): 131101, 网络出版: 2019-07-11   

基于深度网络模型的视频序列中异常行为的检测方法 下载: 1228次

Method of Detecting Abnormal Behavior in Video Sequences Based on Deep Network Models
作者单位
南京工业大学电气工程与控制科学学院, 江苏 南京 211816
摘要
针对视频序列中的几种异常行为,构建训练模型,对其进行识别。使用卷积神经网络(CNN)进行特征提取并采用Adam算法(一种基于梯度的优化算法)进行优化。引入自适应池化层,筛选出判别的特征信息,减轻网络的计算量,加快识别视频序列中存在的异常行为。使用Adam算法对模型进行优化后,识别率可以达到87.6%,引入自适应池化层后,识别率可以达到91.9%。该卷积神经网络对视频序列中基本的异常行为的检测效果比改进的轨迹跟踪(iDT)和双流网络更快更准确;相较于时间分割网络(TSN)和时间关系网络(TRN),识别的准确率稍低,但是识别的速度更快。
Abstract
In this study, a training model was constructed to identify several abnormal behaviors in video sequences. A convolutional neural network (CNN) was used to extract features, and the features were then optimized using a gradient-based optimization algorithm known as Adam algorithm. The adaptive pooling layer was introduced for feature discrimination to reduce the computational complexity of the network and rapidly identify abnormal behaviors in video sequences. The recognition rate reaches 87.6% after using the Adam algorithm for model optimization. The recognition rate reaches 91.9% when the adaptive pooling layer is introduced. CNN is faster and more accurate than the improved dense trajectories and the two-stream networks in detecting abnormal behaviors in video sequences. Compared with the temporal segment networks and temporal relation networks, the CNN has a lower accuracy but a faster speed.

吴沛佶, 梅雪, 何毅, 袁申强. 基于深度网络模型的视频序列中异常行为的检测方法[J]. 激光与光电子学进展, 2019, 56(13): 131101. Peiji Wu, Xue Mei, Yi He, Shenqiang Yuan. Method of Detecting Abnormal Behavior in Video Sequences Based on Deep Network Models[J]. Laser & Optoelectronics Progress, 2019, 56(13): 131101.

本文已被 1 篇论文引用
被引统计数据来源于中国光学期刊网
引用该论文: TXT   |   EndNote

相关论文

加载中...

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!