图像质量评价的建模与仿真研究  

Modeling and Simulation of Image Quality Evaluation

作　　者：李凌[1] 刘志武[1] 李永昌[1] 

机构地区：[1]辽宁工程技术大学应用技术学院,辽宁阜新123000

出　　处：《计算机仿真》2012年第3期284-287,共4页Computer Simulation

摘　　要：研究图像质量准确评价问题,图像采集、压缩传输中会出现畸变,使图像模糊,传统评价方法难以从对这些特征进行正确区分,导致图像质量评价准确率低。为提高图像质量评价准确率,提出一种采用最小二乘支持向量机(LSSVM)的图像质量评价方法。首先采用PSNR和SSIM分别对图像质量进行评价,得到的评价值作为描述图像质量的参数,然后输入到LSS-VM进行学习,建立新的图像质量分类器,采用建立的分类器对图像质量进行仿真评价。仿真结果表明,相对于单一的图像质量评价方法,提高了图像质量评价的准确率,评价结果与视觉感知评估值更加一致。For the distortions often occur in image acquisition,compression and transmission,it is difficult for traditional evaluation method to carry out image quality evaluation.In order to improve the evaluation accuracy of image quality,this paper presented a image quality evaluation method based on least squares support vector machine（LSSVM）.The PSNR and SIMM were used respectively to evaluate the image quality.Then the evaluation values were used as image quality parameters description and input into the LSSVM for learning to establish a new image quality classifier.Finally,the classifier was used to evaluate the image quality evaluation.The simulation results show that,compared with the single image quality evaluation method,the proposed method can improve the evaluation accuracy of image quality,it is more consistent with the evaluation result and visual perception assess value.

关 键 词：图像质量 最小二乘支持向量机 峰值信噪比 结构相似法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]