基于八邻域的自适应P-Laplace图像修复算法

Adaptive P-Laplace Image Inpainting Algorithm Based on Eight Neighbors

针对传统的变分图像修复算法仅仅利用四邻域信息导致的图像过渡不自然、人工干预强等问题,提出一种基于八邻域的自适应P-Laplace修复算法。该算法将待修复像素点的八邻域分为两组四邻域,利用每组四邻域信息,采用中心格式差分方法对待修复像素点进行处理,通过自适应方法确定每组邻域修复的最佳p值,最后再进行加权平均。仿真实验显示,该算法与近几年提出的修复算法相比图像评价较好,PSNR值提高了约5.74%~22.15%,TSSIM值提高约4~30个百分点。

To solve the problems of traditional variation inpainting algorithms caused by simply utilizing four neighbors,like unnatural image transition and strong manual intervention,an adaptive P-Laplace algorithm is devised based on eight neighbors.Firstly,it divides eight neighbors into two four-neighbors,and discretizes points to be repaired by central format difference in each four-neighbor.Then,determines the best p value of each four-neighbor with the adaptive method.Finally,weights and averages two four-neighbors.The simulation experiment results show that we get better PSNR and TSSIM values through the algorithm proposed in this paper,compared with other algorithms in recent years.And the PSNR values are increased by 5.74%~ 22.15%.The TSSIM values are increased by 4~30 percentage points.