Comparing image quality of print-on-demand books and photobooks from web-based vendors

bstract. Because of the emergence of e-commerce and developents in print engines designed for economical output of very short uns, there are increased business opportunities and consumer opions for print-on-demand books and photobooks. The current state f these printing modes allows for direct uploading of book files via he web, printing on nonoffset printers, and distributing by standard arcel or mail delivery services. The goal of this research is to asess the image quality of print-on-demand books and photobooks roduced by various Web-based vendors and to identify correlations etween psychophysical results and objective metrics. Six vendors ere identified for one-off (single-copy) print-on-demand books, and even vendors were identified for photobooks. Participants rank orered overall quality of a subset of individual pages from each book, here the pages included text, photographs, or a combination of the wo. Observers also reported overall quality ratings and price estiates for the bound books. Objective metrics of color gamut, color ccuracy, accuracy of International Color Consortium profile usage, ye-weighted root mean square L*, and cascaded modulation transer acutance were obtained and compared to the observer reponses. We introduce some new methods for normalizing data as ell as for strengthening the statistical significance of the results. ur approach includes the use of latent mixed-effect models. We ound statistically significant correlation with overall image quality nd some of the spatial metrics, but correlations between psychohysical results and other objective metrics were weak or nonexistnt. Strong correlation was found between psychophysical results of verall quality assessment and estimated price associated with