一种基于人眼视觉特性的立体图像质量客观评价方法

摘要

Through simulating contrast sensitivity function, multi-channel effects and stereoscopic perception of human visual system, a quality assessment method of stereoscopic images based on human visual system is proposed. In the quality evaluation of left and right images by computing Canberra distance, wavelet transform was used to simulate multi-channel effects and wavelet coefficients of different spatial frequency were weighted according to the contrast sensitivity function. The proposed method calculated similarity of the absolute disparity images of the original and test stereoscopic images to assess the stereoscopic perception. Then regression analysis was used to integrate the two evaluation results into an equation as our quality assessment model of stereoscopic images. Experimental results show that the proposed objective method achieves consistent stereoscopic image quality evaluation results with subjective assessment.%通过模拟人眼视觉特性中的对比度敏感函数、多通道效应以及立体感知特性,提出了一种基于人眼视觉特性的立体图像质量客观评价方法.在评价左右图像质量时,利用小波变换模拟人眼视觉特性中的多通道效应,不同空间频带的小波系数按对比度敏感函数进行加权,左右图像质量度量采用Canberra $巨离.在评价立体感知时,则通过计算原始与测试左右图像的绝对差值图相似度来实现.其后,通过回归分析将左右图像质量和立体感知评价结果拟合成为所需的立体图像质量客观评价模型.实验结果表明该模型与主观评价结果具有较好的一致性.