中文摘要：

形状局部相似度是衡量三维模型表面两个局部区域之间相似性程度的一个指标，该指标在计算机图形学和计算机视觉领域得到了广泛的应用。基于Spin图，提出了三维形状表面局部相似性的一种度量方法，并将局部相似性度量应用到模型表面的视觉增强应用中。为了能够有效比较模型表面两个局部区域之间的相似性程度，首先对模型上每个顶点邻域沿均匀分布采样方向进行均匀加密采样，然后建立采样点邻域的Spin图作为其形状描述符，通过比较局部区域之间的形状描述符得到局部相似性度量。同时，根据形状的局部相似性度量，对模型表面进行基于相似度的着色，实现模型表面的视觉增强操作。实验表明，基于Spin图的相似度分析方法能够较好地刻画模型表面的相似性程度，方便地实现模型表面的视觉增强效果。

英文摘要：

Local shape similarity, an indicator to measure how similar a region of 3D shape is or dissimilar to another region, has been widely applied to computer graphics and computer vision. Different from traditional curvature map, a novel spin image based on local shape similarity measure is presented in this paper and its application on visual enhancement of 3D models is also given. To efficiently compare two different regions, the neighboring points for each surface vertex are firstly obtained by uniformly sampling along evenly distributed directions on the tangent plane. The spin images are constructed for these uniformly distributed sampling points and the local shape similarity measure can thus be calculated by comparing two spin images of different regions. Finally, due to our proposed local shape similarity definition, an efficient visual enhancement scheme is provided by incorporating our similarity measure into the color adjustment operation. Experimental results indicate that our spin image based on local shape similarity definition is robust and also contributes to visual enhancement.