Optimization of the active contour snaxels for human gait tracking
Mohammadreza Pourfard, Karim Faez, Mohammadjavad Abdollahifard · 2009
Tracking an object is one the most considerable and applicable field of the machine vision. In this field, tracking a non-rigid object like human are so difficult because of variable shape of this kind of object and the self-occlusion that normally between different parts of the human body occurs. The methods of human tracking are categorized to four main categories which are model-based, feature-based, region-based and active contour. Active contour like the other methods has its own advantages and disadvantages and it uses in its own territory because of its speediness and independency of background movements. In previous methods for human tracking through active contour, every snaxels has a limited and constant distance to the others and in some application like human gaits tracking, this is not optimum. In this paper we have proposed a new method to optimize the location of the determined snaxels of the active contour for tracking a human gait and we have improved the speediness and accuracy of it.