基于K-SVD算法的数字图像自适应修复方法  

Digital image adaptive restoration method based on K⁃SVD algorithm

作　　者：王彦龙[1] 高俊杰[1] 杨阳[1] WANG Yanong;GAO Junjie;YANG Yang(Shanxi University,Taiyuan 030006,China)

机构地区：[1]山西大学,山西太原030006

出　　处：《现代电子技术》2024年第13期15-18,共4页Modern Electronics Technique

基　　金：山西省高等学校科技创新项目资助(2022L009)。

摘　　要：为了提升数字图像的完整性和清晰度,提出一种基于K-SVD算法的数字图像自适应修复方法。通过FCM算法将数字图像划分成不同的图像块,将不同类别的数字图像依据K-SVD算法的稀疏编码和字典更新模块进行训练,获取各个不同类别数字图像块的字典,求出其稀疏系数,结合字典和稀疏系数更新数字图像中的每一类图像块,完成数字图像中每一类图像块的修复或重构,将修复好的图像块放回原数字图像中,实现数字图像的自适应修复。实验结果表明,该方法能够有效地恢复图像的细节和结构,修复后的数字图像均方根误差低,并且具有较高的峰值信噪比,同时,修复后的数字图像与原图像的结构相似性高达0.95,且在数字图像修复效率方面具备显著优势。In order to improve the integrity and clarity of digital images,a digital image adaptive restoration method based on K-SVD(K-singular value decomposition)algorithm is proposed.By the FCM(fuzzy c-means)algorithm,the digital image is divided into different image blocks,digital images of different categories are trained based on the sparse encoding and dictionary update module of the K-SVD algorithm,so as to obtain the dictionaries of different categories of digital image blocks,and calculate their sparse coefficients.In combination with the dictionaries and sparse coefficients,each type of image blocks in the digital images are updated,repaired and reconstructed.The repaired image blocks are put back into the original digital image to achieve adaptive restoration of the digital image.The experimental results show that the proposed method can restore the details and structure of the image effectively.The repaired digital image has low root-mean-square error(RMSE)and a high peak signal-to-noise ratio(PSNR).At the same time,the structural similarity between the repaired digital image and the original image is as high as 0.95,and it has significant advantages in the efficiency of digital image restoration.

关 键 词：FCM算法 K-SVD算法 稀疏编码 更新字典 数字图像 图像细节 图像聚类 图像修复 

分 类 号：TN911.73-34[电子电信—通信与信息系统] TP391[电子电信—信息与通信工程]