Transcription of CartoonGAN: Generative Adversarial Networks for Photo ...
{{id}} {{{paragraph}}}
Example: bankruptcy
CartoonGAN: Generative Adversarial Networks for Photo ...
CartoonGAN: Generative Adversarial Networks for Photo CartoonizationYang ChenTsinghua University, LaiCardiff University, Liu Tsinghua University, this paper, we propose a solution to transforming pho-tos of real-world scenes into cartoon style images, which isvaluable and challenging in computer vision and computergraphics. Our solution belongs to learning based methods,which have recently become popular to stylize images inartistic forms such as painting. However, existing meth-ods do not produce satisfactory results for cartoonization,due to the fact that (1) cartoon styles have unique charac-teristics with high level simplification and abstraction, and(2) cartoon images tend to have clear edges, smooth colorshading and relatively simple textures, which exhibit signif-icant challenges for texture-descriptor-based loss functionsused in existing methods.
2.1. Nonphotorealistic rendering (NPR) Many NPR algorithms have been developed, either au-tomatically or semi-automatically, to mimic specific artis-tic styles including cartoons [25]. Some works render 3D shapes in simple shading, which creates cartoon-like ef-fect [28]. Such techniques called cel shading can save
Tags:
Information
Domain:
Source:
Link to this page:
Please notify us if you found a problem with this document: