基于神经网络的人体特征曲线预测方法研究

Research on Prediction Method of Human Characteristic Curve Based on Neural Network

人体特征曲线在人体三维建模中有着非常重要的作用,通常需要手工测量的方式来获取,具有一定的局限性。本文利用神经网络来建立二维图像与三维截面特征曲线之间的关系,并使用三维扫描获取的人体模型来训练神经网络,训练后的神经网络可实现从用户照片直接恢复三维截面特征曲线,并用于三维重建。实验结果表明该方法的训练误差低于2%,能够准确的恢复人体截面曲线特征。

Human characteristic curve plays a very important role in 3 d modeling of human body, which is usually acquired by manual measurement, with certain limitations. In this paper, the neural network is used to establish the mapping relationship between two-dimensional image and three-dimensional feature curve, and the human model obtained by three-dimensional scanning is used to train the neural network. The trained neural network can directly recover the three-dimensional feature curve from the user’s photo and be used for three-dimensional reconstruction. Experimental results show that the error of the method is within 2%, and it can accurately restore the curve characteristics of the real human body.