计算机技术与发展2017，Vol.27Issue(9)：22-25,4.DOI:10.3969/j.issn.1673-629X.2017.09.005
超级计算机的图像结构相似度并行处理方法
Research on Image SSIM Paralleling Algorithm with Super Computer
摘要
Abstract
With the exponential growth of image information,image processing is faced with the dual challenge of low inefficiencies and poor quality. Image Structural Similarity ( SSIM) ,taken as a common method in the field of image quality assessment,is also faced with the identical problem. Therefore,a method of image SSIM parallel processing based on supercomputers is proposed. The image SSIM par-allel processing is implemented based on High Performance Supercomputer Center with computing resources. The three kinds of image in-formation are in the parallel calculation by blocking. The performance of the parallel processing is simulated compared with the serial cal-culation. The experimental results indicate that in allusion to massive image data,it is far lower in the operation time than serial calculation with larger accelerating ratio and higher efficiency,which is effective and feasible,and makes the real-time processing of image possible.关键词
数字图像/超级计算机/结构相似度/加速比Key words
digital image/supercomputing/SSIM/accelerating ratio分类
信息技术与安全科学引用本文复制引用
陈聪梅,都政,井革新,赵晓惠,靳绍巍,陈远磊,饶青雷,邬鸿..超级计算机的图像结构相似度并行处理方法[J].计算机技术与发展,2017,27(9):22-25,4.基金项目
国家重点研发计划(2016YFB0201401) （2016YFB0201401）
