Special Issue Description


Authors : Moiz A. Hussain and G. U. Kharat

Page Nos : 16-22

Description :
An algorithmic approach to detect human motion for video surveillance system is discussed in this paper. The considerable study in this area has been encouraged by the fact that many application areas, including surveillance, Human–Computer Interaction and automatic annotation, will gain from a robust solution. In this paper, a framework for the human motion tracking using two algorithms is presented, first, using AdaBoost classifier. The algorithm consists of three steps: at first, a multiscale image features extracted from an image is used. At the second stage, extracted features are represented by the sparse matrix representation. These sparse matrix represented values are classified using AdaBoost classifier for motion tracking applications. Second using optical flow method where optical flow values are extracted which is then processed by illumination insensitive tracker. Experimental result shows that both the method gives better performance as compared to the other state of the arts methods of video surveillance. Keywords—Video surveliannce; Sparse Matrix; AdaBoost Classifier.

Date of Online: 30 Special Issue-7, Nov. 2015