No-reference stereoscopic image quality assessment based on binocular energy response
Wang Shan-sha · Journal of Optoelectronics·laser · 2014
No-reference image quality assessment is an important task in image processing related applications.In order to achieve quality evaluation for the distorted stereoscopic images with different types of distortion,a no-reference stereoscopic image quality assessment method is proposed based on binocular energy response.In the method,local features are first obtained by applying Gabor filtering on different frequencies,directions and disparities for the distorted stereoscopic images.Then,these local features are fused as stereoscopic features.Finally,support vector regression(SVR)is performed to predict the objective scores by establishing the relationship between the stereoscopic features and the subjective scores.Experimental results show that compared with full-reference and no-reference methods,the Pearson linear correlation coefficient(PLCC)index reaches 0.92,and the Spearman rank order correlation coefficient(SROCC)index reaches 0.93for NBU-3Ddatabase while the PLCC index reaches 0.96,and the SROCC index reaches 0.96for LIVE-3Ddatabase,respectively,which indicate that the proposed method can achieve higher consistency with subjective assessment of stereoscopic images.