基于有约束最小二乘法的IQA算法曲线拟合  

Curve Fitting Based on Constrained Least Squared Method for Image Quality Assessment Algorithms

作　　者：王学珍[1] 刘昱[1] 李器宇[1] 汪少初[1] 

机构地区：[1]天津大学电子信息工程学院,天津300072

出　　处：《电视技术》2012年第1期141-144,共4页Video Engineering

摘　　要：为了实现准确评估图像质量评价算法的目标,采纳VQEG报告中提到的性能指标,完成了将算法数据的曲线拟合过程转换为有约束最小二乘法形式的数学模型,然后编写Matlab程序求解。以PSNR、SSIM和MSSIM三种图像质量评价算法为例的评估结果表明,此法结果准确,求解效率高。in order to evaluate image quality assessment algorithms correctly, the evaluation metrics from the report of VQEG is used, and the curve fitting problem mentioned in that report is transferred into a typical mathematical model of constrained least squared method. Then the fitting curve is a- chieved using Matlab software. Results for the three algorithms of PSNR, SSIM and MSSIM are reasonable, and the fitting procedure is efficient.

关 键 词：有约束最小二乘法 曲线拟合 图像质量评价 MATLAB 

分 类 号：TN911.73[电子电信—通信与信息系统]