Abstract: Based on the total variation methods of gray-scale image inpainting, this paper presents four steps to achieve a color image inpainting: It labels the images to be inpainted, obtains the mask image by labeled pixels, expands the mask image, inpaints the images with the expanding information. Experimental results on related images show that the inpainting result is very well, color image with noise can get good denoised result and keep clear border.

Key words: image inpainting, color image, Total Variation(TV), mask image

摘要： 基于灰度图像修复的整体变分法，提出解决矢量图像修复处理的4步法：对待修补图像进行标注，根据标注的像素灰度值求掩模图像，对掩模图像进行膨胀，根据膨胀信息进行修复。对3幅彩色图像进行实验，取得了很好的修复效果，对含有噪声的图像，也能很好地去噪并保持图像的清晰边界。

关键词: 图像修复, 彩色图像, 整体变分, 掩模图像

CLC Number: 

• TP391