基于Bayes-MeTiS网格划分的3D几何重构

3D Geometric Reconstruction Based on Bayes-MeTiS Mesh Partition

为提升3D模型几何重构过程的压缩效率,提出一种基于MeTiS网格划分的贝叶斯3D模型几何重构算法。首先,在编码端采用MeTiS方法对原始3D网格进行子网划分,采用随机线性矩阵对子网几何形状进行编码,并对边界节点的邻居节点使用伪随机数生成器进行数据序列构建;然后,利用贝叶斯算法进行几何模型重构算法的设计,在理论上给出了均值、方差矩阵以及模型参数学习规则,实现了3D模型的几何重构;最后,将其与图傅里叶光谱压缩(GFT)、最小二乘压缩(LMS)和基于压缩感知的图傅里叶光谱压缩(CSGFT)等算法进行仿真对比。结果表明,所提方法具有较高的比特率压缩指标以及较低的重构误差,计算效率明显提高。

In order to improve the compression efficiency of the geometric reconstruction process of 3 D model,this paper proposed a bayesian geometric reconstruction algorithm based on MeTiS mesh partition for 3 D model.At the encoding part,the MeTiS method is used to realize the subnetting for original 3 Dgrid,the random linear matrix is used to encode the geometry of the subnet,and the pseudo random number generator is used for data sequence construction by considering the neighbor nodes of the boundary nodes.Then,the Bayesian algorithm is used to design the geometric model reconstruction algorithm,and the mean,variance matrix and the model parameters are given in theory to realize the geometric reconstruction of the 3 Dmodel.Finally,by comparing with graph Fourier transform spectral compression（GFT）,least square compression（LMS）and compressed sensing based graph Fourier transform spectral compression algorithms（CSGFT）,the simulation results show that the proposed method has relatively high bit rate compression index and low reconstruction error.