人体模型建模中全局映射智能估计人体形状的方法

Intelligent Estimation Method of Human Body Shape in Human Body Modeling

针对人体模型建模中形状估计存在的性能低、适用性不强等问题,提出一种利用卷积神经网络进行人体形状估计的智能方法,并利用卷积神经网络层次化特征提取和多视图融合的方法精确估计人体形状。以人体模型为基础,对人体三维形状进行尺寸估计,实现快速人体形状估计,且使估计过程自动化,提高估计结果的精确性。研究表明,该方法克服了传统人工测量技术无法捕捉人体特征、估计人体形状与姿态的局限性,可用于人体模型尺寸的测量、服装虚拟现实中人体模型的建模等。

To solve the problems of low performance and weak applicability in shape estimation in human body modeling,an intelligent method using convolutional neural networks for human body shape estimation was proposed,and the principle of convolutional neural network framework for human body shape estimation was explored.The convolutional neural network was used for hierarchical feature extraction and multi-view fusion to estimate human shape accurately.Based on the human body model,size estimation of the three-dimensional shape of the human body was carried out to achieve fast human body shape estimation,and the estimation process was automated,resulting in relatively accurate estimation results.Research shows that this method overcomes the limitations of traditional manual measurement techniques in capturing human features,estimating human shape and posture.The results can be used for measuring human model dimensions,modeling human models in clothing virtual reality,and more.