基于二维人脸图像的三维建模研究  

On 3D Modeling Based on the Two-dimensional Face Image

作　　者：张怡[1] 周渊[2] 

机构地区：[1]兰州工业学院软件工程学院,甘肃兰州730050 [2]兰州交通大学化学与生物工程学院,甘肃兰州730070

出　　处：《兰州工业学院学报》2013年第4期1-5,共5页Journal of Lanzhou Institute of Technology

摘　　要：针对目前三维人脸建模方法存在的不足,在分析了网格重采样方法的基础上,提出了一种改进的基于模板的网格重采样方法.首先对人脸的重要特征点大致定位,自动将人脸划分为36个分块,初步建立三维人脸间的对应关系;然后使用重采样的方法,得到具有规则拓扑结构的人脸特征点,建立较为精细的中性人脸模型.经实验证明,该方法能够在较短时间内实现对中性三维人脸的建模,同时对高维模型进行了网格简化.In view of the shortcomings of current 3D face modeling method and the analysis of the grid resampling method,an improved grid resampling method based on template is put forward.Firstly,the important feature of the human face is roughly positioned and automatically divided into 36 sub-blocks to establish three-dimensional correspondence;Then use resampling methods to get regular topology facial feature and establish a more sophisticated neutral face model.The experiment proved that this method can achieve three-dimensional face modeling in a relatively short period of time,and the high-dimensional model is simplified.

关 键 词：三维人脸建模 形变模型 特征点提取 重采样 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]