基于三维配准的面部变化检测技术

Facial change detection methods based on 3d registration

针对微整容变化检测的问题,提出一种基于改进迭代最近点算法的变化检测方法。首先采用面部数据曲率特征确定特征点,然后通过普式分析法对面部特征点集求解最小二乘函数,提高迭代最近点算法的效率。通过对面部微整形数据的测试,与经典的ICP方法相比,文中方法具有配准精度高、收敛速度快的特点。

In order to solve the problem of micro plastic change detection,this paper presents a change detection method based on improved ICP algorithm. First,the feature points are generated by surface curvature. Second,solving the least squares function of facial feature points by the procrustes method. In the experiments,facial micro plastic data is adopted to do the tests. Compared with the ICP method,the proposed method shows faster convergence and more accurate registration.