Title:	EVENT DETECTION IN SURVEILLANCE VIDEOS OF PUBLIC PLACES
Author(s):	Mina Shirvani Boroujeni, Mohammad Rahmati
ISBN:	978-972-8939-74-8
Editors:	Yingcai Xiao
Year:	2012
Edition:	Single
Keywords:	Visual surveillance, event detection, abandoned luggage, tracking, occlusion handling, finite state machine
Type:	Full Paper
First Page:	83
Last Page:	90
Language:	English
Cover:
Full Contents:	click to dowload
Paper Abstract:	Nowadays, video surveillance systems are widely used in public places to increase the security, and thus the intelligent visual surveillance systems, capable of detecting potentially dangerous situations, play important roles in such systems. In this paper we propose an event detection system, to detect an abandoned luggage in surveillance videos, which represents a potential threat to public safety. The proposed system consists of four major modules. The first module, foreground segmentation, detects foreground regions corresponding to the objects present in the scene. The second module, stationary object detection, detects a suspicious luggage item and its owner, who are the main targets in a luggage abandonment event. The owner is then tracked by the third module, the moving object tracking module, which can effectively detect and handle complex object interactions and occlusions in real world situations. Finally the forth module, the event detection module, analysis the owners trajectory and set alarm in case of detecting an abandoned luggage item. The proposed system is evaluated on PETS 2006 dataset and the results are promising.
	Go Back