基于双树复数小波的结构相似度测量法

以人类视觉系统更符合提取图像结构信息为前提的结构相似度评价模型，因其高效特性受到广泛重视，但与传统图像质量评价方法一样，其性能易受比较图像几何变化的影响.为了有效地评价各类图像质量，提出了一种双树复数小波与结构相似性测量相结合的全参考客观评价方法.基于双树复数小波的平移不变性、方向选择性和低冗余特性，首先对图像进行双树复数小波变换，并在复数域计算结构相似度，然后经平均获得客观质量评价分，改善了对几何失真图像的评价性能，且实现简单、计算量小.在LIVE图库上的实验结果表明，提出的模型对亮度、对比度以及几何变化具备更强的稳定性.

Structural similarity measurement based on dual-tree complex wavelet

Structural similarity index,which is based on the assumption that the human visual system is highly adapted for extracting structural information from the scene,has attracted wide attention for its efficiency.But its performance,the same as that of traditional image quality assessment methods,will be liable to geometric distortions.This work proposed a full-reference method based on dual-tree complex wavelet transform(DTCWT) and structural similarity in order to efficiently assess the image quality.Through using DTCWT's properties of shift invariance,directional selectivity and low redundancy,a structural similarity measurement model in dual-tree complex domain was established,which is simple and low computational complexity.First,the image is processed by using DTCWT,and their structural similarity scores are computed in complex domain,then the objective quality assessment score is obtained via mean.The results of the experiment on the LIVE database demonstrate the model's stabilization to different types of distortions such as luminance,contrast and small geometric transform.