Color Characterization Of Desktop Inkjet Printers

For a desktop color printer, which is a typical output device, the characterization consists of forward and inverse models. The forward model seeks to convert the device-dependent control values (e.g., RGB) to the device-independent reflectance spectrum or colorimetric stimulus. The inverse model, which is more useful in color applications, transforms from a requested reflectance spectrum to the optimal control values that, when printed, could reproduce that spectrum or colorimetric stimulus with greatest possible accuracy.First of all, establish the forward model, based on the principle of physical mixed color. The method investigates the printer color separation and medium color rendering processes, based on which, spectral reflectance can be predicted. The YNSN and enhanced YNSN models, which are based on simple and complicated dot gain respectively, are employed in the printer color separation process. As the color separation cannot be directly controlled, the colorimetric error objective function is designed. The mapping between RGB and CMYK is modeled using polynomial regression, and is resolved by nonlinear optimization. In the paper, we also use 3D-LUT and local polynomial regression to establish the forward model. Finally, six field local space algorithm is presented. Through the weight parameters and the optimization of power factor, the algorithm precision is improved.For the inverse model, first establish the mapping between CIEXYZ and RGB with polynomial regression, which is under a single light. The algorithm has good colorimetric accuracy under the light and leads to bad results in a different light. To solve the problem, the method of RGB averaging is presented. Finally, a local space numerical optimization method is presented. Through the optimization of parameters, the algorithm precision is improved and the good accuracy of the print of multispectral image is realized.