A hybrid snake for selective contour detection
Juan Zapata, R. Ruı́z · International Conference on Signal Processing · 2007
A hybrid snake algorithm is presented in this paper. The procedure is based on combining a dynamic balloon force, a GVF force and a traditional gradient force. The balloon force is modified in order to increase the capture range and convergence speed of active contour in a dynamic way. Our interest is to give the balloon force a bigger preponderance at the early of procedure and a smaller weight at the later stage. The superposition of effects get through the GVF force and the gradient force is ruled by a user parameter. Our GVF force is computed for a decreased number of iterations during the calculation process and so the diffusion field is located near edges only. In this manner, we can use this snake for segmentation of objects with strong edges inside or outside. The synthetic experiments and real-time face tracking application presented in this paper prove the feasibility of this hybrid snake.