| 基于仿射变换和线性回归的3D人脸姿态估计方法 |
| |
| 引用本文： | 邱丽梅,胡步发. 基于仿射变换和线性回归的3D人脸姿态估计方法[J]. 计算机应用, 2006, 26(12): 2877-2879 |
| |
| 作者姓名： | 邱丽梅 胡步发 |
| |
| 作者单位： | 福州大学,机械工程及自动化学院,福建,福州,350002;福州大学,机械工程及自动化学院,福建,福州,350002 |
| |
| 摘 要： | 提出了一种由仿射变换关系到线性回归的3D人脸空间姿态估计方法。即跟踪到人脸特征点后，根据仿射变换关系得到人脸姿态的粗估计值，以这个粗估计值作为人脸姿态的初始值，再通过线性回归迭代求得人脸姿态的精确值。实验结果表明，该方法在较大的姿态变化范围内，具有良好的估计精确度和鲁棒性。
|
| 关 键 词： | 人脸姿态估计 仿射变换 线性回归 |
| 文章编号： | 1001-9081（2006）12-2877-03 |
| 收稿时间： | 2006-06-23 |
| 修稿时间： | 2006-06-232006-09-04 |
3D face pose estimation based on affine transformation and linear regression |
| |
| QIU Li-mei,HU Bu-fa. 3D face pose estimation based on affine transformation and linear regression[J]. Journal of Computer Applications, 2006, 26(12): 2877-2879 |
| |
| Authors: | QIU Li-mei HU Bu-fa |
| |
| Affiliation: | College of Mechanical Engineering and Automation, Fuzhou University, Fuzhou Fujian 350002, China |
| |
| Abstract: | A method for estimating 3D space face pose was proposed based on affine transformation and linear regression. Namely, after tracking face feature points, a roughly estimated value was obtained based on affine transformation relationship, and then this rough value was taken as the starting value, precise value of face pose was obtained based on linear regression iteration. Finally, the experimental results show that the method achieves better estimation accuracy and robustness in a wide range of face poses. |
| |
| Keywords: | estimation of face pose affine transformation linear regression |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《计算机应用》浏览原始摘要信息 |
|
点击此处可从《计算机应用》下载全文 |
|
