Modeling of growth via active contour models

The authors analyze the growth of biological specimens with the help of Active Contour Models known as Snakes. Conventional shape-change models like optical flow, deformable templates, etc., primarily address rigid and non-rigid deformations which are not in the form of growth. Moreover the above models assume the shape-change between successive frames in an image sequence to be small which is not true in many real-world biomedical image sequences. The Active Contour Model presented here tracks the boundary of an object as it changes shape and can also predict how the object might change shape. The algorithm is an energy minimization technique which tracks the growth of a specimen by attracting a closed piecewise cubic curve to the boundary of the specimen in every image of a sequence. This method computes shape-change especially when it is in the form of growth-better than techniques like optical flow, deformable templates, etc. Unlike conventional shape-change algorithms, the authors´ algorithm can be applied iteratively to compute drastic shape-changes also. The authors also propose a parallel framework to facilitate real-time implementation