Editable 3D Model Based on Generative Adversarial Network.

In recent years, 3D modeling technology has become more and more popular in emerging VR and AR applications. Due to the complexity and diversity of hair styles in the real world, hair modeling is still challenging, so it is very important to study hair modeling methods. This paper presents an interactive single - view 3D hair generation method based on generative adversarial network. We construct a parameter transformation by generative adversarial network to map the 2D hair to the 3D hair structure. The new 3D hair is generated by editing the 2D hair information and the 3D hair interactive operation is realized. Interactive editing mode mainly includes drawing mode. The results show that this method has good flexibility and superiority in 3d hair generation and user controllability.