Abstract: To address the LBF model’s dependence on the initial contour and the weak control over the edge, we investigated a method of the level set image segmentation combining saliency and edge information. First, we generated the image saliency map combined with the theory of wavelet analysis and based on the visual attention mechanism. Then, the edge detection function was constructed using wavelet decomposition information. It was introduced into the improved LBF level set model together with the adaptive initial contour, and the finite difference method was used for numerical solution. The experimental results show that the image segmentation method proposed in this paper can effectively reduce the influence of the initial contour position on the active contour model, and can yield better segmentation results for both synthetic images and natural images. Compared with other traditional methods, it is of higher evolution efficiency and segmentation quality.  

Key words:  , image segmentation, wavelet decomposition, saliency map, adaptive, level set