An Image Segmentation Algorithm Based on Watershed and Snake Model
Yinlong Wang, Qianjin Li · 2018
The Snake model is widely used in the fields of edge extraction, image segmentation, etc. The model can initialize the target appropriately and make the convergence of the target, so that the energy is in the minimum state to achieve the goal separation effect. When the initial position of the target is sensitive, it needs to rely on other mechanisms to reasonably initialize the internal energy. Due to the non-convexity of the model, it is possible to converge to the local extreme point or even divergence. We apply the watershed algorithm to the energy segmentation algorithm of active contour model The improved watershed algorithm is used to determine the initial contour of the active contour model. The iterative method is used to retrieve the local approximation points around the contour points to select the smaller contour model. When the minimum value is obtained, the target contour is extracted.