中文摘要：

为了有效、实时地对各种类型失真立体图像质量进行评价,提出了一种基于极端学习（ELM）和四元数小波交换（QWT）的无参考（NR）立体图像质量评价方法。首先利用SSIM密度立体匹配模型生成相关的视差图、差异度可信图和右视图差异补偿图3D映射图;然后分别对左右视图、视差图和差异度可信图进行QWT,计算图像QWT第3相位系数相位幅值加权标准差和能量;再计算右视图差异补偿图统计特征熵和中值;最后将所提取的所有特征输入到基于核映射ELM学习,预测失真立体图像质量。在LIVE3D图像质量评价数据库上的实验结果表明,本方法与人类主观质量评分具有较好的一致性。在LIVE 3D图像质量库I（Phase I）和库II（Phase II）上的斯皮尔曼相关系数（SROCC）分别达到0.926和0.914。

英文摘要：

To assess stereoscopic image quality with various types of distortion effectively, a no-reference （NR） stereoscopic image quality assessment is proposed which uses extreme learning machine （ELM） with kernel to learn the features based on image quaternion wavelet transform （QWT）. Firstly, it adopts the left view image and the right view image of stereoscopic image as the input of the structural similari- ty （SSIM） based dense stereo matching algorithm to obtain 3D perceptual maps of the stereoscopic im- age. an estimated disparity map, an estimated disparity confidence map and an estimated disparity-com- pensated fight view image. Secondly, it processes the left view image, the right view image, the estimated disparity map, the estimated disparity-compensated right view image with QWT. Thirdly,it computes the energy and weight standard deviation of all image QWT coefficient. Then it computes the statistical fea- ture entropy and median of the estimated disparity confidence map. At last, these features above are used as the input of ELM with kernel to predict the quality of the tested stereo images. The experiment is based on no-cross image material in LIVE 3D image quality database, and the Spearman rank ordered correlation coefficients （SROCCs） are 0. 926 in phase I and 0. 914 in phase II, which indicates proposed method is consistent with subject quality assesment.