摘要
在研究Criminisi算法的基础上,针对Criminisi算法修复顺序不可靠、采用固定大小的样本块模板尺寸以及在修复过程中容易产生误差累积等缺点,提出了一种改进的图像修复算法,该算法结合改进的优先项函数、根据梯度信息自适应选择模板大小等方法。实验结果表明,该算法易于实现,与Criminisi算法相比,能够取得更好的修复效果。
On the basis of researching Criminisi algorithm, to unreliable repair order, adopting fixed sample template size and easily producing cumulative error over the process of the repair in Criminisi algorithm, an improved image restoration algorithm is proposed. The algorithm consists in the improved priority function, choosing adaptively template size according to gradient information. The experimental results show that the algorithm is easy to implement, and can get better repair effect compared with Criminisi algorithm.
出处
《电子设计工程》
2014年第23期174-176,180,共4页
Electronic Design Engineering
关键词
图像修复
优先权
模板窗口
置信度
梯度
image restoration
priority
template window
confidence item
gradient
