中文摘要：

提出一种用于构造给定三维模型的拟合Loop细分曲面的迭代优化算法,使得拟合曲面与原始模型之间的逼近误差最小.算法中的逼近误差定义为原始模型各面元到拟合曲面最小距离的积分.与Loop细分小波分解算法的比较表明,该算法以适度的运行时间代价得到了更优的结果.此外,该算法还可以加以推广,作为一类从输入模型生成其近似表示的优化算法的基础.

英文摘要：

This paper presents an iterative algorithm to construct a fitting Loop subdivision surface from a given 3D model by minimizing the approximation error between the fitting surface and the original model. The approximation error is defined as the integral of the minimum distance between each surfel on the original model and the fitting surface. By comparing this algorithm with the Loop subdivision wavelet decomposition algorithm, it is shown that our algorithm generates better results at a moderate running time cost. Moreover, our algorithm can be extended to serve as the basis of a class of optimization algorithms for constructing the approximation representation from an input model.