A GAN-based approach toward architectural line drawing colorization prototyping

Abstract Line drawing with colorization is a popular art format and tool for architectural illustration. The goal of this research toward generating high-quality natural-looking based on an line drawing. This paper presents new Generative Adversarial Network (GAN)-based method, named ArchGANs , including ArchColGAN ArchShdGAN . GAN-based line-feature-aware network stylized generation. lighting effects generation network, from which the building depiction in 3D can benefit. In particular, able to maintain important features correlation property parts as well reduce uneven caused by sparse lines. Moreover, we proposed color enhancement method further improve Besides single images, also extend our handle image sequences achieve rotation animation. Experiments studies demonstrate effectiveness usefulness prototyping.