An evaluation model for character quality on scanned image

Scanners, standalone or embedded in MFP, have been widely utilized for document scanning/copying in office environments. There are many factors affecting the quality of scanned image. They include quality of characters, size of gamut, degree of noises, color reproduction, tone characteristics, and gray balance, etc. This paper is focused on quantitative evaluation of character quality on scanned images. Objective of this paper is to derive a quantitative evaluation model that faithfully matches with the human perception. Key attributes determining character quality are extracted first based on subjective human visual experiments. For each of the identified attributes, a quantitative metric is developed. The proposed evaluation model is designed as a linear combination of the quantified metrics. Coefficients for linear combination are estimated by applying linear regression. Various experiments are performed to verify the performance of the proposed evaluation model.