基于人类视觉的感知立体图像质量评价方法  

A perceptual stereoscopic image quality assessment method based on human visual system

作　　者：佟晓煦 李素梅[1] 刘富岩[1] 孟迪[1] 徐姝宁 

机构地区：[1]天津大学电子信息工程学院,天津300072

出　　处：《光电子．激光》2015年第2期378-386,共9页Journal of Optoelectronics·Laser

基　　金：国家"863"高技术(2012AA011505;2012AA03A301);国家自然科学基金重点基金(60932007);国家自然科学基金(61002028)资助项目

摘　　要：为了实现对不同失真类型的立体图像进行质量评价,提出了一种基于人类视觉的立体图像质量客观评价方法,分别从图像清晰度与立体感两方面进行评价。图像清晰度方面,将原始与失真立体图像分解为5个带通图像后利用对比度敏感度函数(CSF)优化各失真带通图像,并模拟掩盖效应,通过整合各原始带通图像,综合感知误差,构造信噪比(SNR)作为评价图像清晰度的性能指标;立体感方面,对绝对差值图像进行视觉感知模拟,建立SNR指标评价立体感的优劣。实验结果表明,对不同失真类型立体图像的评价结果表明,Pearson线性相关系数(PLCC)与Spearman等级相关系数(SRCC)均优于现有评价方法。In order to achieve quality assessment for the distorted stereo images with different types of distortions, a stereo image quality objective assessment method based on human visual system （HVS） is proposed in this paper. This method assesses the stereo image in two aspects,image fidelity and stereo sense. In view of the image quality, the original and distorted stereo images are divided into five bandpass images. The contrast sensitivity function plays an important role in optimizing the distorted bandpass im- ages and masking effect is simulated. The original bandpass images and the perceptive errors are synthe- sized and the signal-to-noise ratio as the performance index of evaluating stereo image quality. As for the stereo sense, this paper focuses on the absolute disparity images and simulates them. At last, the signal- to-noise ratio is calculated. The experimental results show that the Pearson linear correlation coefficient （PLCC） index and the Spearman rank correction coefficient （SRCC） index by using the proposed meth- od are both superior to existing assessment methods in the assessment results for stereo images with dif- ferent types of distortions.

关 键 词：立体图像 质量评价 立体感 人类视觉系统(HVS) 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]