基于2D头部图像的3D建模及优化  

3D Head Modeling and Optimization Based on 2D Images

作　　者：宋重钢[1] 高鑫[1] 孙光泽[1] 李旭[1] 李龙泉[1] 李涛[1] 

机构地区：[1]南开大学信息技术科学学院,天津300071

出　　处：《南开大学学报（自然科学版）》2011年第6期57-61,共5页Acta Scientiarum Naturalium Universitatis Nankaiensis

基　　金：南开大学本科生创新科研项目(BX8-158);中央高校基本科研业务费专项资金

摘　　要：提出了一种基于2D头部图像进行3D重构的方法.通过光线补偿、肤色建模以及对图像进行膨胀和去离散点等技术,从2D正交头部图像中提取头部特征点,在标准人体头部模型上通过适配与修改完成3D头部重构,并采用3D渲染使头部模型更具立体感和真实感,并可实现360°观察,在刑侦、饰物佩戴、发型选择等方面具有重要应用.A method is proposed as three-dimensional head modeling and optimization based on two-dimensional images.Feature points on the human head are gathered through light compensation,skin color modeling and the procedure of inflation and corrosion.By modifying the standard human head model,the reconstruction of 3D human head model is accomplished.The 3D render could enable the model be observed from all aspects.The results demonstrate that this methodology could reconstruct the 3D model of a human head efficiently,which is encouraged to be utilized in various domains like detection and online shopping.

关 键 词：光线补偿 肤色建模 特征提取 3D渲染 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]