Abstract: Shape-Text matching is an important task of high-level shape understanding. Current methods mainly represent a 3D shape as multiple 2D rendered views, which obviously can not be understood ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results