基于流形调和变换的浅浮雕生成算法

针对利用3D网格生成浅浮雕不能使用精确的网格表示、浮雕细节保持特性依赖于规则采样高度域上分辨率的问题,提出一种基于3D网格频域分析的浅浮雕算法.从信号处理的角度出发,首先利用网格的几何微分性质和拓扑结构通过离散调和变换将网格的深度信号变换到频域；其次提出一种基于几何频率能量的频域划分策略,将频率分割为低频、高频和噪声,再对低频进行线性压缩,利用各项同性的扩散函数对高频进行自适应非线性压缩并过滤噪声,得到信号在频域上一组新的变换系数；最后利用调和逆变换将处理后的频域映射到信号的空域重建浮雕模型.实验结果表明,该算法可通过控制频谱的分割和扩散函数的参数实现浮雕模型细节的增强.

Bas-relief Generation Using Manifold Harmonics Analysis

In this paper a novel algorithm for bas-relief generation from 3D shape is presented based on manifold harmonics spectral analysis.The algorithm focuses on achieving detail-preserving bas-relief independent on the resolution of height field by using exact mesh representation.Firstly,the coordinates of the mesh vertices can be seen as discrete signal,and are transformed to the frequency domain by means of discrete harmonic analysis.During the transformation,the differential geometry properties and topology information of the mesh are utilized.Secondly,according to the proposed definition of energy of harmonic transform coefficients,the frequency is segmented into low,high and noisy frequency bands.Subsequently,the algorithm ignores the noisy frequency component,compresses the low frequency component linearly into a small range,and nonlinearly compresses the high frequency adaptively.Finally,the new frequency is mapped back to space domain by use of reverse harmonic.Experiments demonstrate that the detail on the relief can be enhanced by tuning parameters of an isotropic diffusion filter and using the strategy of spectral segmentation.