Towards Comprehensive Understanding of Event Detection and Video Summarization Approaches

Abstract

Intelligent video surveillance system plays a vital role in ensuring safety and security to the public. In recent years, surveillance systems are widely used in all places starting from border security application to street monitoring systems. As the analysis of surveillance video becomes one of the most emerging trends and research fields of interest, there is a need for a comprehensive survey on various works being carried out in the field. A single surveillance camera records the happenings in a particular field of view both day and night at all time, likewise multi-camera environment generates a huge volume of data. Though motion sensitivity based cameras considerably reduce the storage volume, they still produce large volume of data when employed in busy areas. The large amount of video data cannot be fully observed for analysis as it is time overwhelming process. Hence it is required to summarize the activities in the scene and detect unusual events recorded in videos. Hence, this paper discusses various techniques for summarization and unusual event detection in video and it presents current scenario of research in this area. This paper lists out the main focus of various researchers, datasets being used and features being considered for their approaches. It yields the future directions in this area of research, applications of event detection and summarization.