Structure analysis image quality measurement

In this paper, a novel image quality assessment, called Structure Analysis Image Quality Measurement (SAIM), is proposed by evaluating the difference of the distorted image from the reference image on three aspects: the average luminance, edge intensity and orientation. It is sensitive to edge change, which consists well with the perceptual characteristics of human vision. The experiment results demonstrate its superiority over the most popular metrics, such as PSNR and MSSIM, on “LIVE Image Quality Assessment Database”. And furthermore, SAIM can find a wide application in various fields, because of its capability of estimating the similarity between any two signals.