Virtual3D Brush Model Based On Elastic Cone

Chinese Calligraphy is a treasure of eastern traditional art, which attracts many computer graphics researchers and e-artists. Consequently, the realistic simulation of Chinese Calligraphy is an important part of e-art, including hairy brush simulation and ink diffusion simulation. The simulation methods of hairy brush are roughly categorized into three classes:2D brush model,3D brush model and data-driven modeling approach,2D simulation can render stroke in real time, but performance is poor;3D simulation approach generates brush footprints based on brush deformation, which has good performance, but computation is high; Data-driven modeling approach simulate the constrained dynamics of brushes by using a small lookup table of real brush deformations. The approach can’t recreate every possible brush deformation due to the restricted set of brush deformations. The ink diffusion simulation is mainly based on particle diffusion.This paper proposes a new3D brush model based on elastic cone. We regard a single bristle of brush as an elastic rod and the brush bundle absorbing ink as an elastic cone. We use discrete elastic rod theory to compute deformation of the brush when writing on a paper, and the brush footprint is generated by the intersection region between the deformed brush and the paper plane. To efficiently produce smooth natural stroke, we introduce interpolation and texture mapping between two adjacent footprints and skeleton-based stroke generation method. To mimic ink diffusion in real time, we give a simple method based on texture mapping and gray attenuation:using gray gradual exponentially attenuates form the center to border of the stroke.We present a simple data-driven method to simulate the splitting of a brush, which records sub-line quantity, width and interval between adjacent sub-lines of every splitting, and build a lookup table of the splitting. This method selects ideal splitting type using lookup table when rendering splitting stroke.Experimental results show that our method do run fast and is suitable for creating realistic calligraphy.