摘要
基于Perona-Malik的各向异性扩散提出曲率驱动的图像修补模型,探讨并利用可加算子分裂方法实现快速求解。所提模型通过选择适当的参数,使得扩散强度能更好地适应图像细节变化。实验结果显示了较满意的图像修补效果。
A model of inpaniting by curvature-driven was proposed based on Perona-Malik＇ s anisotropic diffusion, and fast solution schemes were exploited by applying AOS methods. In this model, diffusion strength can adapt to the changes of image details by selecting suitable parameters. The experiments show this inpainting method is effective and acceptable.
出处
《实验室研究与探索》
CAS
2008年第2期19-22,共4页
Research and Exploration In Laboratory
