基于最佳匹配块搜索算法的皮质性白内障OCT图像自动修复方法研究

Research on the automatic restoration method of cortical cataract OCT image based on the best matching block search algorithm

常规方法填充皮质性白内障OCT图像待修复区域结构信息时,轮廓信息提取过程复杂,填充不完整,导致修复图像信息熵较低,修复精度较差。针对这一问题,提出基于最佳匹配块搜索算法的皮质性白内障OCT图像自动修复方法。划分图像已知区域和待修复区域,计算区域边界像素点的修复优先权,选取优先权最高像素点为中心,设置修补目标块,搜索对应的最佳匹配块,通过最佳匹配块搜索算法,提取匹配块颜色信息,利用带通滤波器,调节扩展因子和移动因子,提取轮廓信息和细节信息,填充至待修复区域,完成破损像素点修复。选取医院提供的原始白内障OCT图像序列,设置对比实验,结果表明,设计方法相比设计对比方法,修复后图像信息熵提高了2.51 dB、2.89 dB,修复精度提高了1.04%和1.22%,充分保证了图像修复质量。

When the conventional method fills in the structural information of the area to be repaired in the OCT image of cortical cataract,the contour information extraction process is complicated and the filling is incomplete,resulting in lower information entropy of the repaired image and poor repair accuracy.Aiming at this problem,an automatic restoration method for cortical cataract OCT images based on the best matching block search algorithm is proposed.the known area of the image and the area is divided to be repaired,the repair priority of the pixels is calculated at the boundary of the area,the highest priority pixel is selected as the center,the repair target block is set to search for the corresponding best matching block,and the best matching block is used as search algorithm.The color information of the matching block is extracted,use the band-pass filter,adjust the expansion factor and the movement factor,extract the contour information and detail information,fill in the area to be repaired,and complete the repair of the damaged pixel.Select the original cataract OCT image sequence provided by the hospital and set up a comparison experiment.The results show that compared with the design comparison method,the image information entropy after restoration is increased by 2.51dB and 2.89dB,and the restoration accuracy is increased by 1.04%and 1.22%,which is fully guaranteed Improved the quality of image restoration.