作　　者： ; ; ; (姜求平）;

机构地区： 宁波大学信息科学与工程学院

出　　处： 《光电子．激光》 2013年第12期2409-2415,共7页

摘　　要： 立体图像质量评价是评价立体视频系统性能的有效途径,而如何利用人类视觉特性对立体图像质量进行有效评价是目前的研究难点。本文提出了一种基于视差空间图(DSI)的立体图像质量客观评价方法。首先,分别构造原始立体图像和失真立体图像的DSI图;然后,通过三维离散余弦变换(3D-DCT)提取出反映图像质量和深度感知的特征信息,并采用主成分分析(PCA)进行特征降维,形成立体图像特征信息;最后,通过支持向量回归(SVR)建立立体图像特征与主观评价值的关系,从而预测得到立体图像质量的客观评价值。实验表明,对于对称立体图像库,Pearson线性相关系数(PLCC)和Spearman等级相关系数(SROCC)值均达到0.94以上;对于非对称立体图像库,PLCC和SROCC值分别达到0.94和0.91以上。结果表明,本文方法能够很好地预测人眼对立体图像的主观感知。 Stereoscopic image quality assessment is an effective way to evaluate the performance of stereoscopic video system. However,bow to utilize human visual characteristics in quality assessment is still an issue. In this paper,an objective stereoscopic image quality assessment method based on disparity space image （DSI） is proposed. In this method,we first construct DSIs from the original and distorted stereoscopic images, respectively. Then, three-dimensional discrete cosine transformation （3D-DCT） and principal component analysis are used to extract the stereoscopic image feature information from the DSIs that reflect the image quality and depth perception. Finally,support vector regression （SVR） is performed to predict the objective scores of stereoscopic images by establishing the relationship between the stereoscopic image feature information and the subjective scores. Experimental results show that compared with other methods, the Pearson linear correlation coefficient （PLCC） and Spearman rank-order correlation coefficient （SROCC） indicators reach 0.94 on symmetric stereoscopic image database,and the PLCC indicator reaches 0. 94 and the SROCC indicator reaches 0. 91 on asymmetric stereoscopic image database,which indicate the proposed method can achieve higher consistency with subjective assessment of stereoscopic images for both symmetric and asymmetric databases.

关 键 词： 立体图像质量评价 视差空间图 三维离散余弦变换 特征降维 支持向量回归

领　　域： [电子电信] [电子电信]