基于能量信息耦合梯度调节机制的图像修复算法设计与应用

摘要:

为了克服当前图像修复算法主要依靠图像的置信度信息来获取优先修复块，忽略了图像的能量信息，导致修复结果中存在不连续及伪吉布斯现象等缺陷。本文设计了基于能量信息与梯度调节机制的图像修复算法。首先，通过区域能量函数来求取图像的能量信息，以计算待修复块的优先权信息，得到优先修复块。然后，基于图像梯度模值，建立梯度调节机制，以调节样本块的大小，获取与图像纹理相适应的样本块尺寸。引入平方差求和函数，以确定最优匹配块。最后，通过像素点间的差异性，构造相似惩罚因子，以更新置信度项，完成图像的修复。实验结果显示，较当前图像修复方案而言，所提算法具备更好的修复性能，所得到的修复图像拥有更好的纹理连贯性与更高的结构相似值。

Abstract:

In order to overcome the current image restoration algorithms mainly rely on the image confidence information to obtain the priority repair block, ignoring the image energy information, making the algorithm"s repair performance decline, resulting in the repair image discontinuity and pseudo Gibbs phenomenon and other defects. In this paper, energy information and gradient adjustment mechanism are used to repair the image. Firstly, the energy information of the image is obtained by the region energy function, and the priority information of the block to be repaired is calculated by the data and confidence terms. Then, on the basis of image gradient modulus, a gradient adjustment mechanism is established to adjust the size of the sample block and obtain the sample block size corresponding to the image texture. Finally, the sum function of square difference is introduced to calculate the similarity between the block to be repaired and the matching block, so as to obtain the optimal matching block. Through the difference between pixels, construct the similarity penalty factor to update the confidence term and complete the image restoration. The experimental results show that the algorithm in this paper has better performance than the current algorithm, better texture coherence and good structural similarity.