首页 | 本学科首页   官方微博 | 高级检索  
     检索      

具有网络一致结构的三维人脸模型重建研究
引用本文:刘 政,董洪伟,杨 振.具有网络一致结构的三维人脸模型重建研究[J].教育技术导刊,2009,19(10):156-159.
作者姓名:刘 政  董洪伟  杨 振
作者单位:江南大学 物联网工程学院,江苏 无锡 214122
摘    要:针对传统方法及机器学习方法对大量三维人脸数据、训练样本数量与质量依赖性大的问题,采用基于光照立体的方法,利用人脸图像重建三维人脸。综合利用基于法线的模型变形法和非刚性变形法,提出一种基于法线的非刚性变形算法。利用SFS算法计算顶点法线,然后使用法线和局部刚性约束使参考模型变形,进而得到与参考模型具有一致网络结构的三维人脸模型,从而得到三维人脸图像。该方法与传统方法相比,节省了大量样本要求,并且立体效果更好。

关 键 词:SFS  三维人脸  非刚性网格变形  
收稿时间:2020-03-28

Research on 3D Face Model Reconstruction with Network Consistent Structure
LIU Zheng,DONG Hong-wei,YANG Zhen.Research on 3D Face Model Reconstruction with Network Consistent Structure[J].Introduction of Educational Technology,2009,19(10):156-159.
Authors:LIU Zheng  DONG Hong-wei  YANG Zhen
Institution:School of Internet of Things Engineering, Jiangnan University, Wuxi 214122, China
Abstract:Aiming at the problem that traditional methods and machine learning methods depend on a large number of 3D face data, the number and quality of training samples, this paper uses the method based on lighting stereo to realize the reconstruction of a three-dimensional face from a face image. A non-rigid model deformation algorithm based on normal is proposed, which combines a normal based deformation algorithm and a non-rigid model deformation algorithm. The SFS is used to estimate the normal of the vertices, then updated normal are used to deform the reference face model with local rigid constrain, finally a deformed 3D face model is obtained, which has the same mesh structure with the reference model, thereby obtaining a 3D face image. Compared with the traditional method, the proposed method can save a lot of sample requirementsand the stereo effect is better.
Keywords:SFS  3D face  non-rigid mesh deformation  
点击此处可从《教育技术导刊》浏览原始摘要信息
点击此处可从《教育技术导刊》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号