Application of PQS for Image Quality Analysis in Visible Spectral Imaging

摘要

An image quality investigation in visible spectral imaging was performed. Spectral images were simulated using different number of imaging channels, wavelength steps, and noise levels based on practical spectral imaging systems. A mean opinion score (MOS) was determined from a subjective visual assessment scale experiment for image quality of spectral images rendered to a three-channel display. A set of partial image distortion measures, including color difference for color images, were defined based on classified and quantified actual distortions produced by spectral imaging systems. Principal components analysis was then carried out to quantify the correlation between distortion factors. Finally, a multiple regression analysis (MRA) was carried out between the principal component vectors and the measured MOS values to determine the picture quality scale (PQS). The obtained quality metric, PQS, had high correlation with the subjective measure, MOS. The importance of contribution of the distortion factors in the image quality metric was also evaluated.