中文摘要：

提出一种考虑方向场的体纹理合成新方法．在二维纹理样本中搜索邻域，采用优化的邻域匹配算法逐个合成体素；在匹配二维纹理样本像素邻域和体素邻域时，考虑各体素上的方向，生成各向异性体纹理，并允许采用各向异性的二维纹理样本获得满意的合成结果．方向场可由过程式方法生成，也可由三维医学数据处理而得．实验结果表明，从动物的弥散张量图像得到的方向场能有效地合成真实的动物肌肉纹理，为计算机图形学提供了一种新的肌肉建模手段．

英文摘要：

We present a novel approach for vector field guided solid texture synthesis from 2D examples. One distinctive feature of our algorithm is that the guidance of vector field is incorporated into the neighborhood searching and color optimization for each voxel, hence it is capable of synthesizing anisotropic solid texture to conform with the input vector field. The guidance vector field can be procedurally generated, or computed from 3D medical images. We show that biologically meaningful muscle can be conveniently modeled via the vector field derived from diffusion tensor images of real animal muscles.