CAJ | 학술논문

为了提高融合图像质量评价中评价结果与人眼视觉特性的一致性，分析了现有融合图像质量评价方法，提出了一种基于图像结构信息复数表示的融合图像质量评价方法，通过计算图像亮度分量的梯度，构成了一种表征图像结构信息的梯度复数矩阵，用该矩阵表征图像的结构信息。考虑到复数无法计算互信息等参数，将分块奇异值分解后得到的矩阵作为度量矩阵，采用该矩阵计算了两种融合图像质量评价方法。实验结果表明，该方法提高了评价结果与人眼视觉特性的一致性，对于融合效果较好的金字塔和小波方法给出了3.7485和3.7222的评价结果，与人眼视觉特性的一致性优于传统方法。
위료제고융합도상질량평개중평개결과여인안시각특성적일치성，분석료현유융합도상질량평개방법，제출료일충기우도상결구신식복수표시적융합도상질량평개방법，통과계산도상량도분량적제도，구성료일충표정도상결구신식적제도복수구진，용해구진표정도상적결구신식。고필도복수무법계산호신식등삼수，장분괴기이치분해후득도적구진작위도량구진，채용해구진계산료량충융합도상질량평개방법。실험결과표명，해방법제고료평개결과여인안시각특성적일치성，대우융합효과교호적금자탑화소파방법급출료3.7485화3.7222적평개결과，여인안시각특성적일치성우우전통방법。
In order to improve the consistency of the assessment result of image fusion with that of Human Vis -ual System , the state-of-the-art image fusion assessment methods are deeply analysed , then a new assessment method is proposed in this paper , which is based on the complex number expression for image structure .The gradient information of luminance layer of color image is used to perform the task .When it is used to describe image structure , more human visual system-sensitive information are contain in the corresponding complex ma-trix.Due to the calculation problem of mutual information , we perform singular value decomposition on the complex matrix, and the singular value vector of each image block is used to construct the new matrix .Results from experiments show that the proposed method gives evaluation of 3.748 5 and 3.722 2 for pyramid and DWT methods .It improves the consistency of assessment results with those of human visual system .