一种真实感三维人脸交互式生成方法

An Interactive Approach to Generate Realistic 3D Face

针对传统三维人脸模型构建复杂、效率低下等不足,提出了一种基于单张照片的真实感人脸快速生成方法。首先,构建标准化的人脸模型库,交互式选取输入照片的面部特征点,并基于脸型特征,匹配最佳模型;其次,通过三角形变形和双线性插值的方法实现照片到三维模型的纹理映射,并引入Alpha图实现人脸覆盖纹理到模型中性纹理的融合过渡;最后,采用网格调节方法从整体到细节分层次调整模型以达到生成真实感三维人脸的目的。实验结果证明,该方法对用户友好,所需特征点较少,可实时生成高质量的真实感模型。

As traditional 3D face construction is too much complex and has low efficiency, therefore, an interactive approach is proposed to generate realistic 3D faces based on one photo. Firstly, we con- structed a standardized face model database, and interactively select the facial feature points from an input picture, and then, the best-matched face model is selected based on the features. Secondly, a triangular deformation approach and bilinear interpolation are used to achieve texture mapping from a photo to 3D model. The Alpha Fig was introduced to realize a fusion transition from the photo texture to the neutral texture. At last, a model regulator to regulate the model hierarchically was designed to generate realistic 3D faces. Experimental results show that this approach is user-friendly, and can generate a high-quality realistic character in real time with less characteristic points.