基于高分辨灰度掩模的图像修复改进设计

Improved Design of Image Repair Based on High Resolution Gray Mask

图像修复是结合计算机图形学和虚拟现实技术的综合学科,在文物保存与防护、医学图像资料与影视特效这些方面具备许多的使用意义.以前的图像修复算法采用断点标定和特征提取方法,当出现不连续破损断点时,图像修复效果不好.提出一种基于高分辨灰度掩模的图像修复改进算法.首先进行破损待修复图像的边缘检测和灰度特征提取,提取一个和目前等待修正回复模块中优先级最高的最好样本模块,然后通过计算得出损坏等待修正恢复图像中的低频系数向量和高频系数向量,进而得出图像特点的分层提升成果,通过Mallat小波提升方案设计,基于高分辨灰度掩模处理,跟踪破损图像的破损区域走向,提高对破损图像修复的全局搜索能力.仿真结果表明,该算法的对图像修复质量较高,信噪比误差较小,体现了其优越性.

Image restoration is combined with computer graphics and virtual reality technology comprehensive discipline. In the cultural relics preservation and protection, medical image information and video effects the use of these areas have many meanings. Previous image restoration algorithm using breakpoint calibration and feature extraction method, when there is a discontinuous broken breakpoint, image inpainting result is bad. Based on a high resolution gray mask image restoration algorithm. First damaged to repair the image edge detection and gray feature extraction, extraction and currently waiting for a reply correction module in the highest priority of best sample module, and then calculated the damage wait for correction coefficient of low frequency and high frequency coefficient vector of image vector, image characteristics of layered ascension results are obtained, by Mallat wavelet lifting scheme design, based on high resolution gray mask processing, tracking to the breakage of the damaged image area, improve the global search ability of damaged image restoration. The simulation results show that the algorithm of image restoration quality is higher, signal--to--noise ratio error is small, embodies its superiority.