Research On Face Image Vectorization Based On Cartoonish Sketch Editor

Image vectorization technology in science and engineering has a wide range of applications, such as image compression and transmission, construction, electronics engineering and so on. Because of the complexity of the image itself, although Image vectorization technology have decades of research and development, the various vectorization algorithm does have some limitations. Today, the image vector algorithm is still evolving, this technology is still has a broad research space.With the rapid development of cartoon industry, the productions of the quality and the production efficiency have an increasingly higher requirement. Comics technology platform which using vector-based technology, has a significant effect on quality and production efficiency of cartoon production.This article around the bitmap image vectorization technology research and development, based on the cartoonish sketch editor, describes the bitmap image to a vector of basic theory and methods, as well as color image processing related theories and methods. On this basis, the human facial images vectorization method is proposed, the main content of this article include the following:1. Comprehensive study of image processing mathematical morphology and it's expanding in gray mathematical morphology and color mathematical morphology.2. Based on the color mathematical morphology approach, using multi-structural elements, combined with the results of the regional division expunges the redundant information, and then the contours of face images obtained.3. Analyze the skin-color's clustering in the YCbCr color space, and establish the skin-color model in the YCbCr color space, and then detection the image using this skin-color model, after color segmentation, denoising, and regional verification steps, finally obtain the face region.4. According to the prior knowledge of facial features, combined with the human eye and mouth features in the skin color model, locate the facial features and coding the feature region, using Bezier curves and circle or arc fitting the face of the feature profile, namely, using curve fitting method to achieve vectorization.