The Contour Extraction Based On Skin Model And Cartoon Stylized System’s Design And Implementation

In recent years, with the increased demand for applications of image processingsoftware, image processing softwares are more and more, such as PhotoShop, Picasa.Image processing software is becoming an essential product in the daily life of manypeople. People could landscape photos through this kind of software. So I want toachieve a image processing system which can extract the portraits contours andcartoon-style function. The contours of the effect is similar to the sketch, which justdon’t have the effect of the brush thickness. Cartoon stylized photos can meetpeople’s portrait landscaping purposes.In this paper, we first complete the function of the contour extraction. Based onthe color image segmentation technique, we study and compare the characteristics andprinciples of the commonly used color space, the final choice is doing the imagesegmentation in the YCbCr color space. Then we study the extraction of imagecontours algorithm, compare the efficiency and ease of them, the final choice is theinternal hollowed contour extraction method. I am not very satisfied with the featuresof the initial results map, so we summed up the shortcomings in the algorithm. On thisbasis, we do the details’ optimization based on the refined ASM feature locationalgorithm, using local threshold instead of the global threshold for imagesegmentation. Experimental results show that the improved method better. Then wehave studied the image stylization algorithm, and finally select the style method basedon the adaptive bilateral filtering. In order to achieve portrait style features, followedby the color space conversion, abstract filtering, edge detection, brightnessquantization, IBW sharpening functions. And this system is a way to interact with theuser, the user can manually configure some of the parameters, and then select theirsatisfaction with the results.