Title :
Distraction Free Evolution of Active Contours
Author :
Srikrishnan, V. ; Chaudhuri, Subhasis
Author_Institution :
Electr. Eng. Dept., Indian Inst. of Technol. Bombay, Mumbai
Abstract :
We propose a novel energy term to make the curve evolution quite robust to spurious edges. The physical intuition behind the formulation is that an object edge is generally continuous, but it could be composed of weak and strong segments. We then formulate the energy term which depends on a second order measure defined on the contour. Minimisation of this energy term yields a space varying curvature based curve evolution equation. An added advantage of the formulation is that this term also acts as the regularising term for smoothing the curve evolution. The proposed term can therefore be used in conjunction with any of the numerous gradient based active contour models. For our experimentation purpose, we have used the well-known gradient vector force model as the external force. We have performed a number of experiments on images and obtained good results.
Keywords :
edge detection; image segmentation; smoothing methods; curve evolution equation; curve evolution smoothing; distraction free evolution; energy minimisation; gradient based active contour; gradient vector force; object edge; Active contours; Computer vision; Energy measurement; Equations; Image segmentation; Level set; Pattern recognition; Robustness; Smoothing methods; Active Contour; Gradient based segmentation; Salient contour;
Conference_Titel :
Advances in Pattern Recognition, 2009. ICAPR '09. Seventh International Conference on
Conference_Location :
Kolkata
Print_ISBN :
978-1-4244-3335-3
DOI :
10.1109/ICAPR.2009.18