基于三维人体模型的尺寸测量和拟合

Size Measurement and Fitting Based on 3D Human Body Model

提出一种对三维人体模型的躯干相关数据进行测量的方法,以及两种根据宽度值和厚度值计算维度值的回归模型。首先在CAESAR人体模型数据库中选择一个标准体型的人体作为测量模板,利用MeshLab在该模板上选出围绕待测量部位的水平点列,记录每个点坐标以及该点所在的三角网格信息。之后计算点列重心坐标,利用重心坐标系计算数据库中其他人体模型的待测量部位数据,主要的计算对象是胸部和腰部的相关数据。在得到大量的宽度和厚度数据后,分别用线性回归和神经网络进行训练,得到两种拟合函数。最后测量现实中的人体数据,代入到拟合函数中进行准确度验证。

Proposes a method for measuring the torso correlation data of a three-dimensional human body model, and two regression models for calcu? lating the dimension values based on the width value and the thickness value. First, selects a standard human body as the measurement template in the CAESAR human body model database, uses the MeshLab to select the horizontal points around the part to be measured, re? cords the coordinates of each point and the triangle mesh information of the point. Then calculates the barycentric coordinates of each point, and uses the barycentric coordinate system to calculate the data of the part to be measured of other human body models in the data? base. The main calculation objects are related data of chest and waist. After obtaining a large amount of width and thickness data, the two regression functions are obtained by linear regression and neural network respectively. Finally, the actual human body data is measured and substituted into the fitting function for accuracy verification.