A Review on Multiple Objects Tracking in a Video Scene with Particle Filtering Techniques

摘要

Multiple objects tracking in a video scene is one of the most challenging tasks in the field of computer vision as well as it is highly demand in many computer applications, such as surveillance, tracking of animals, pedestrians, vehicles, human behavior analysis, and so on. A complex scene is characterized by several moving objects such as people, animals, vehicles, etc. In order to perform higher level tasks, such as to fight against terrorism, crime, public safety for efficient management of traffic, etc. which are demanded by typical surveillance or monitoring applications, to detect, identify and track various objects of interest. There are many approaches have been proposed to solve this problem, it still remains challenging due to factors like abrupt appearance changes and severe object occlusions. The objectives of multiple object tracking are to place moving objects in sequential video frames. In this paper, we contribute the first comprehensive and most recent review on this problem using Particle filters. We provide a discussion about issues of multiple objects tracking via Particle filters research, as well as some interesting directions which could possibly become potential research effort in the future.