Analysis of Abnormal Behaviors in Specific Scenarios Based on SSD

Because the current troubleshooting of various abnormal behaviors in surveillance video requires a lot of manpower and can’t be processed in time, this paper proposes abnormal behaviors analysis in a specific scenario based on SSD, so that the surveillance camera can detect and recognize the abnormal behavior of the object in real time. The algorithm in this paper is applied to the surveillance video scene of the general hotel front desk. The SSD backbone network is used to extract the convolution feature and the average pool feature of the dataset, and then multi-scale classification of feature maps of certain feature layers and regression, and finally through the NMS processing output algorithm finally detected the object’s confidence and coordinate. Experiments show that the algorithm of our proposed algorithm is close to 90%, and the processing speed reaches 15FPS, which basically meets the real-time detection of abnormal human behaviors in specific scenes in surveillance video.

[1]  Lianfen Huang,et al.  Real-time detection algorithm of abnormal behavior in crowds based on Gaussian mixture model , 2017, 2017 12th International Conference on Computer Science and Education (ICCSE).

[2]  Bill Triggs,et al.  Histograms of oriented gradients for human detection , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).

[3]  Kaiming He,et al.  Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks , 2015, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[4]  Wei Liu,et al.  SSD: Single Shot MultiBox Detector , 2015, ECCV.

[5]  Tony Lindeberg,et al.  Scale Invariant Feature Transform , 2012, Scholarpedia.

[6]  Ali Farhadi,et al.  You Only Look Once: Unified, Real-Time Object Detection , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[7]  Feng Niu,et al.  HMM-Based Segmentation and Recognition of Human Activities from Video Sequences , 2005, 2005 IEEE International Conference on Multimedia and Expo.

[8]  Jing Cai,et al.  Video crowd detection and abnormal behavior model detection based on machine learning method , 2018, Neural Computing and Applications.

[9]  Jian Zhang,et al.  Face Detection with Effective Feature Extraction , 2010, ACCV.

[10]  Christian Szegedy,et al.  DeepPose: Human Pose Estimation via Deep Neural Networks , 2013, 2014 IEEE Conference on Computer Vision and Pattern Recognition.

[11]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.