An improved skeletonization algorithm of color cerebral vascular image based on color level set mode

This paper presents a continuous skeleton extraction method of color image. This method makes use of a new color level set model based on Bayesian classifier and uses two intermediate functions to extract the topology information and skeleton in HSV space. Not require pre-processing of color images can we get a complex gradient color image skeleton using this method. Because all the parameters are obtained from the analysis, so it avoids artificial intervention. Experiment and analysis can verify the validity and robustness of this method and it is not sensitive to the color gradient and the boundary noise.