摘要
分形图像压缩算法具有较高压缩比,但存在压缩时间长和重构图像质量不高的问题。为解决以上问题,提出一种基于皮尔逊相关系数与核密度估计的低秩稀疏分形图像压缩算法。该算法提取R块(值域块)和D块(定义域块)的皮尔逊相关系数作为特征量,并对提取的皮尔逊相关系数特征量进行最优带宽核密度估计,再利用分形图像编码的低秩稀疏分解实现R块和D块的匹配。将本文算法分别与基本分形编码(basic fractal image coding,BFIC)、稀疏分形图像压缩算法(sparse fractal image coding,SFIC)、双层非负矩阵分解算法(double-layer non-negative matrix factorization,DLNMF)和正交稀疏分形编码算法(orthogonal sparse fractal coding,OSFC)进行比较,实验结果表明,图像的重构质量和编码速度都得到了提高,减少了图像的存储空间和传输带宽,重构后能够保持图像细节,在医疗图像、媒体数据传输、遥感监测等工程领域有较好的应用前景。
Fractal image compression algorithm has high compression ratio,but there are problems of long compression time and low quality of reconstructed images.In order to solve these problems,a low rank sparse fractal image compression algorithm based on Pearson correlation coefficient and kernel density estimation is proposed.The proposed algorithm extracts Pearson correlation coefficient of R block(range block)and D block(domain block)as feature quantity,and the optimal bandwidth kernel density estimation is performed on the extracted Pearson correlation coefficient features,then the low rank sparse decomposition of fractal image coding is used to realize the matching of R blocks and D blocks.The proposed algorithm is compared with basic fractal image coding(BFIC),sparse fractal image coding(SFIC),double-layer non-negative matrix factorization(DLNMF)and orthogonal sparse fractal coding(OSFC).The experimental results show that the image reconstruction quality and coding speed are both improved.The storage space and transmission bandwidth of the image are reduced,and the image details can be maintained after reconstruction.It has a good application prospect in engineering fields such as medical images,media data transmission and remote sensing monitoring.
作者
张琴
谢莹
曹一青
ZHANG Qin;XIE Ying;CAO Yiqing(School of Mechatronics and Information Engineering,Putian University,Putian 351100,China)
出处
《贵州大学学报(自然科学版)》
2025年第3期35-43,共9页
Journal of Guizhou University:Natural Sciences
基金
国家青年科学基金资助项目(62205168)
国家自然科学基金资助项目(62276146)。
关键词
图像压缩
分形编码
皮尔逊相关系数
核密度估计
image compression
fractal coding
Pearson correlation coefficient
kernel density estimation
