Unsupervised deep learning system for local anomaly event detection in crowded scenes
- PDF / 1,854,422 Bytes
- 21 Pages / 439.37 x 666.142 pts Page_size
- 41 Downloads / 195 Views
Unsupervised deep learning system for local anomaly event detection in crowded scenes Anitha Ramchandran 1 & Arun Kumar Sangaiah 1 Received: 29 January 2019 / Revised: 2 April 2019 / Accepted: 26 April 2019 # Springer Science+Business Media, LLC, part of Springer Nature 2019
Abstract Anomaly detection in video surveillance is a significant research subject because of its immense use in real-time applications. These days, open spots like hospitals, traffic areas, airports are monitored by video surveillance cameras. Strange occasions in these recordings have alluded to the anomaly. Unsupervised anomaly detection in the video be endowed with many challenges as there is no exact definition of abnormal events. It varies as for various situations. This paper aims to propose an effective unsupervised deep learning framework for video anomaly detection. Raw image sequences are combined with edge image sequences and given as input to the convolutional auto encoder-ConvLSTM model. Experimental evaluation of the proposed work is performed in three different benchmark datasets such as Avenue, UCSD ped1 and UCSD ped2. The proposed method Hybrid Deep Learning framework for Video Anomaly Detection (HDLVAD) reaches better accuracy compared to existing methods. Investigating video streaming in big data is our further research work. Keywords Video surveillance . Abnormal event detection . Crowd analysis . Convolutional auto encoder . Convolut LSTM
1 Introduction Video analysis is an active research area in computer vision domain. The massive amount of video data is available today because of surveillance cameras installed in almost every part of our society. Growth in hardware technologies and processing power, less cost of surveillance cameras and existing of large video data made video analysis as a trending and significant
* Arun Kumar Sangaiah [email protected] Anitha Ramchandran [email protected]
1
School of Computing Science and Engineering, Vellore Institute of Technology, Vellore, Tamil Nadu 632014, India
Multimedia Tools and Applications
research area [12]. It has many real-time applications such as human behavior recognition, traffic monitoring, and violence detection. In most cases, video surveillance cameras are installed for security concerns. Analyzing gigantic amount of video data is a tedious task. So, intelligent surveillance is highly needed where human operators can be alerted automatically when there is any abnormality in the recorded video. In general, intelligent video analysis research area has two paths, event recognition and anomaly detection. The first one concentrates on interpreting video semantically (e.g. human activity recognition) whereas the second one concentrates on finding unusual or rare events. Video anomaly detection is a meticulous task for intelligent surveillance cameras. It is really important in real-world applications as it is able to capture rare or abnormal events. Angela A Sodemann et al., [32] provided a detailed review of anomaly detection in the surveillance vid
Data Loading...
