Characterization Of Printer Based On The Spectral Neugebauer Model

With the development of computer and copying technology, the accurate reproductionof the variety of color images has become increasingly important. The standardization(characteristics) question of the printer which is one of the important proofing equipments forcolor reproduction is a very important topic.The core technology of the printing device characterization is the color spaceconversion. Taking into account the model-based characterization methods need adequatetheoretical basis and mathematical formulas to support, but only smaller amounts ofmeasurement data, this paper adopts the model-based characterization methods to study thecharacterization question of the printer. At present, Yule-nielsen modified spectral Neugebauermodel is widely used, therefore, this paper is to study the characterization based on theYule-nielsen modified spectral Neugebauer model of printer.Considering that the Yule-Nielsen modified spectrum Negubauer model ischaracterized by the ordinary least square and the total least square method which do notutilize any uncertain information associated with measured spectrum, therefore, the resultingmodels could be too sensitive to inevitable spatial nonuniformities and color drifts that arisedepending on when and where on the paper the measurements are taken. This dissertationestablishment the robust least square method spectral Neugebauer characterization modelbased on analysising the characteristics of the spectral Neugebauer model and the robustestimation theory, and make a comparative study with ordinary least square spectralNeugebauer characterization model.On the basis of deeply analysising the ordinary least square and robust least squaremethod spectral Neugebauer characterization model, we make numerical experimentscomparative study of the corresponding dot area rate and Neugebauer primary color spectralreflectance of the ordinary least square method and the robust least square the spectralNeugebauer characterization model. Meanwhile, from two angles of the spectral reflectanceerror and CIEl976L*a*b*color error, we analysis and contrast the prediction accuracy of thetwo kind of the spectral Neugebauer characterization model. Thus we can obtain thatcompared with the ordinary least square method, the prediction accuracy of the robust leastsquare method spectral Neugebauer characterization model is higher and more stable, able towithstand the gross error impact. And the IGG estimation method spectral Neugebauercharacterization model have higher prediction accuracy and more robust ability than theHuber estimation method spectral Neugebauer characterization model.