A Local √3-Subdivision Algorithm preserving Sharp Features

A local √3-subdivision algorithm to rebuild sharp features is presented. Adaptive subdivision is performed only in the areas facing to the viewpoint in the pre-process. And in each adaptive subdivision, we choose the normal angle Θ of the adjacent faces as the control error to detect whether or not the approximation is sufficient. Sharp edges are automatically tagged according to the given threshold Θ and subdivision masks are designed based on different sharp degrees. Experimental results indicate that the algorithm requires less memory but keeps sharp features better