Artist-Friendly Strip Mesh Hairstyle Modeling System

Hair plays an important role in character molding,so research on simulating and modeling realistic hairstyles has never stopped.Limited by computing capability of the hardware,hair mesh models still dominate in real-time application fields,such as mobile games and social interaction.Due to the high computation complexity of a large number of hair strands,realistic hair models are not widely used currently.However,artists still model hair mesh models by cumbersome handwork,with lacking of convenient systems.As a result,this dissertation creates an artist-friendly strip mesh hairstyle modeling system that automates most of the tedious operations,and improves the efficiency of hair model generation.To reduce the difficulty of hair modeling,we use the idea of decomposition to divide a complete hairstyle into a small number of different regions,and each region generates a mesh model according to the input.At the same time,the framework of the system is extracted from the traditional artistic hair modeling process,that is,the hairstyle is disassembled into two layers of strip-based meshes,the inner layer and the outer layer.We use a contour model to control the shape of hairstyle,and a few curves are used for describing the general direction of the hair.With a little interaction,the geodetic distance-based mesh segmentation algorithm is used to uniformly divide the inner mesh to get strip meshes of inner layer,which is obtained by the contour model.And we implement an iterative generation algorithm of hair strip meshes based on image thinning,which can quickly obtain hair strip meshes of outer layer with few intersected faces from the input.To further improve the effect,we provide stylization and post-processing operations.In addition,editing tools of hair curves allow users to adjust the shape fast.Because the system framework is based on the artists’ actual production process,it can achieve hair effects that better meet the expectations of them.