基于单张人脸图像三维形状模型重建

摘要

Aiming at the request of facial poses in reconstructing 3D face model from a single face image, a new method is proposed to reconstruct 3D face shape model from a single face image. First, the face poses can be obrained from the method of face pose estimation, and the face shape database is created. Then the 3D face shape model is reconstructed by principal component analysis (PCA). Finally the more precise 3D face shape model is obtained using radius base function (RBF). Experimental results show that the proposed method can obtain satisfying reconstruction model, and is more robust with face images with rotated poses and feather points position error.%在人脸图像识别优化的研究中,针对由单张人脸图像重建三维模型时对人脸图像姿态存在要求的问题,为了提高识别精度,提出基于单张人脸图像姿态预估计和主成分分析(PCA)的形状模型重建算法.首先由三维姿态估计方法得到人脸姿态,并建立人脸形状模型样本库,然后通过选取的特征点,利用主成分分析进行三维人脸形状模型的重构,最后利用径向基函数(RBF)变换和特征点坐标精确调整三维人脸形状模型,并进行仿真.仿真结果表明,重构的三维人脸形状模型效果良好,提高了精度,对有旋转姿态的人脸图像和特征点定位误差也有很好的鲁棒性.