Abstract: The exemplar-based image inpainting algorithm, proposed by Criminisi and his partner, is analysed, in view of some deficiency in calculating the patch priority, an improved method for exemplar-based image inpainting is proposed. The original image is decomposed into structure part and texture part by TV-based decomposition algorithm, the patch priority is calculated by using the structure image, which makes the priority computation more precise. Experimental results demonstrate that the proposed method achieves better results for structure-edge inpainting.

Key words: image inpainting, image decomposition, texture inpainting, priority

摘要： 分析了Criminisi等人提出的基于样本的图像修复算法，针对其在计算修复块优先级时存在的一些不足，提出一种改进的基于样本的图像修复方法，通过基于TV模型的分解算法将待修复图像分解为结构图像和纹理图像，利用结构图像来计算修复块的优先级，使得优先级的计算更加准确。实验结果表明该方法对图像结构边缘的修复有明显的改善。

关键词: 图像修复, 图像分解, 纹理修复, 优先级

CLC Number: 

• TP391.41