Research On Interactive Person Cartoonlization Method

The development of the Internet promotes the rapid development of the digital entertainment industry. As an important part, cartoon is popular and impressive in its own way. Currently, the mainstream of generating cartoons by computer is to use algorithms to automatically generate cartoons. However, these methods are less effective and there are many defects, especially for some pictures with occlusion. If we correct the defects generated by human interaction in the process of generating the cartoon, we will be able to get a complete, personalized cartoon. That is to say, we can generate our own interactive cartoon in a convenient way, and also we can replace the cartoon style using the interactive method. This article focuses on the further study of the above problems. We have studied how to generate the corresponding cartoon image for a positive face image through an interactive way, and on this basis, we have proposed a method to generate multi-style cartoon. We can design one’s favorite, complete and personalized cartoon image through human interaction. At the same time, this paper have further studied the method for generating video cartoon, i.e., how to generate the corresponding cartoon video against someone’s video on the basis of the known cartoons synthesis method.The research content is detailed as follows:1. An interactive semiautomatic face cartoon synthesis method was presented to generate a corresponding input face cartoons in an interactive way. The study includes major facial organ generation, interactive eyebrow correction, outline correction, colorize the cartoon image, etc. The generated cartoons are more vivid and have better visual effect.2. In order to generate more personalized complete cartoon image, a method was proposed to generate a multi-style cartoon. We can get a personalized and complete cartoon on the basis of the cartoon synthesis method, including hair replacement module, expression transformation module, glasses addition module and body addition module. The corresponding results of experiments and analysis of algorithms were given, indicating that our method can generate personalized, complete cartoon image for each operation.3. On the basis of the cartoon synthesis method, a method for generating video cartoon was further introduced. Optical flow method is used to track changes in facial organs’ main feature points, so as to get a cartoon video which is the same as the corresponding facial movements.4. The software implementation of the interactive face cartoon synthesis system was also given to complete all the content of the above algorithms. Through the system proposed in this paper, the users can easily make their own cartoon images, and the system is simple to use and has high real-time performance.