基于改进结构相似度的立体图像质量评价方法

摘要

文中提出了一种基于改进结构相似度的立体图像质量客观评价算法.针对结构相似度方法不能较好地评价噪声与模糊失真的立体图像质量的缺陷,文中将结构相似度函数引入频域,并用纹理信息作为权重计算平均结构相似度.实验结果表明,文中方法针对噪声和失真立体图像评价分数与主观评价结果存在较强的相关性.%An objective quality assessment method of stereo image based on improved structural similarity Index( SSIM) is proposed in this paper. In view of the fact that the original method based on structural similarity index fails to assess the quality of noise and fuzzy distortion stereoscopic image, structure similarity function has been introduced into the frequency domain and then the average structural similarity index has been computed based on texture information coefficient. The experimental results show that there is a strong correlation between assessment results from the proposed method and subjective evaluation results.