A Novel Ant Colony Detection Using Multi-Region Histogram for Object Tracking

Seid Miad Zandavi, Feng Sha, Vera Yuk Ying Chung, Zhicheng Lu, Weiming Zhi

Abstract: Efficient object tracking become more popular in video processing domain. In recent years, many researchers have developed excellent models and methods for complicated tracking problems in real environment. Among those approaches, object feature definition is one of the most important component to obtain better accuracy in tracking. In this paper, we propose a novel multi-region feature selection method which defines histogram values of basic areas and random areas (MRH) and combined with continuous ant colony filter detection to represent the original target. The proposed approach also achieves smooth tracking on different video sequences, especially with Motion Blur problem. This approach is designed and tested in MATLAB 2016b environment. The experiment result demonstrates better and faster tracking performance and shows continuous tracking trajectory and competitive outcomes regarding to traditional methods.