摘要
近年来,基于非线性高阶偏微分方程的高质量图像修补算法已经得到了发展,但这些方法需要大量的迭代,时间开销大,复杂度高。Telea提出的基于FMM的修补算法可以快速完成修补,但存在行进方向和边缘信息保持的问题。对此进行了改进,采用MFM方法,并引入扩散张量。实验结果表明提出的方法可以达到较高的质量,而且速度快。
High quality image inpainting methods based on nonlinear higher-order partial differential equations have been developed in the last few years. These methods need a large number of iterations, and result in large time cost and high computational complexity. Inpainting technique based on fast marching method proposed by Telea can complete inpainting just through single pass, but had some problems in marching direction and edge information maintaining. The improvement in this paper is based on MSFM and coherence-enhancing diffusion. Experiment results show that the novel algorithm meets the high quality and is quite fast.
出处
《信息技术》
2008年第10期27-30,共4页
Information Technology