Comprehensive Imaging Quality Assessment Method for Laser Printer Based on SVM

Abstract

Researches on the laser printer imaging quality assessment mainly focus on the independent quality assessment of the lines, dots, characters, images, etc. In this paper, a comprehensive method of imaging quality assessment for laser printers is presented based on support vector machine (SVM). Firstly, test samples are designed. The images of line and character, area and dot acquired from the laser printer are taken as samples. Then, different quality indexes are adopted to evaluate the different aspects of laser printer imaging quality, and the index values of those samples are calculated. The quality indexes are selected according to international standard ISO 13660, including line width error, roughness, ambiguity, darkness of the line and character, gray standard deviation, graininess, mottle of the field area and roundness error of the dot. Finally, the samples are evaluated by subjective assessment method of grade classification, and then the relation model between quality indexes and subjective assessment results are established using SVM. It is proved by experiments that the developed relation model is effective and feasible to predict comprehensive imaging quality results for laser printers, which has good consistency with human vision perception.