摘要
针对古纺织文物模拟修复的应用场合,提出一种基于纹理修复的优化改进算法。该算法利用矩阵值的结构张量准确估计图像中的边缘方向和变化强度,结合Criminisi修补算法,可以在原图和修补区域之间的纹理连接上有更好的连续性。给出了该方法的具体步骤,通过实验证明该方法有更好的视觉效果。
Aiming at the restoration of cultural relics of the ancient textile simulation applications, put forward a kind of improved algorithm based on Optimization of texture repairing. The algorithm using the value matrix of the structure tensor accurate estimation image edge direction and intensity change, criminisi algorithm for repairing the combination, can be connected with a better continuity of texture between the original image and the repair area. Given the specific steps of the method, through experiments prove the method has a better visual effect.
出处
《电脑知识与技术（过刊）》
2016年第9X期193-195,共3页
Computer Knowledge and Technology
关键词
古纺织物
图像修复
Criminisi
结构张量
ancient textile fabrics
image restoration
Criminisi
structure tensor
local matching
