摘要： 传统的全视角人脸纹理图像生成方法计算量大、鲁棒性差，不能很好地解决配准和接缝问题。为此，提出一种改进方法。对待拼接的人脸图像进行归一化处理，利用基于列特征向量的灰度值匹配算法实现图像的精确匹配，使用双重融合算法处理重叠区域，使图像平滑过渡。实验结果表明，该方法匹配速度较快、鲁棒性较强，能有效消除误匹配。

关键词: 三维人脸建模, 图像拼接, 图像融合, 全视角纹理图像

Abstract: Considering the high computational complexity, low robustness, and inefficiency in solving image matching and stitching seam by traditional method for generating global view face texture image, an improved method is proposed in this paper. The face images to be spliced are normalized. An exact match of images is achieved according to gray scale image matching algorithm based on the column and feature vectors. The double fusion algorithm is used to smooth the overlapping area. Experimental results show that this method can eliminate false matches validly and maintain good image stitching accuracy with a high speed.

Key words: 3D face modeling, image stitching, image fusion, global view texture image

中图分类号: 

• TP391