Evaluation of active contour-based tracking methods

This paper presents and evaluates three different approaches to track moving objects from moving camera image sequence using active contour models. The first approach is region-based and makes use of cumulative histograms, whereas the other two approaches are boundary-based and they exploit both b-spline and wavelet representations, respectively. Various performance parameters are studied to the merits and demerits of the three approaches. Different case studies and operational conditions are considered to highlight the significance of each approach. The contribution of this work is to provide suitable identification schemes. Depending on contour-based tracking, the most suitable identification approach could be chosen under various conditions. Copyright © 2009 Inderscience Enterprises Ltd.