基于曲面形变的三维人脸样本配准

3D Face Registration Based on Surface Deformation

三维人脸样本是人脸识别、人脸动画等领域进行模型训练、算法设计及性能比较的重要数据平台。为了提高三维人脸样本的配准效果和配准速度,提出一种基于曲面变形的三维人脸样本配准方法。该方法通过对标准化样本实施一系列的变形操作来建立原始样本与标准样本之间的对应关系。然后基于该对应关系对原始样本进行配准处理,在处理过程中使用统计方法对样本上的毛刺点和空洞进行修补。实验结果表明,使用该方法可以快速、有效地对不同格式的原始样本进行配准。

3Dface database is an important data platform for model training,algorithm design.To improve the matching result and efficiency of 3Dface sample,we proposed a new registration method based on surface deformation.First a series of deform operation is performed on the template sample to get the correspondence between template sample and raw sample.The registration operation is performed on the raw sample.In the procedure of matching process,the statistical method is used to tackle the problem of noise point and hollow.The experiment results show that the proposed method has good performance on face registration.