Optical Models for Appearance Management of Printed Materials - (2016-2019) defended

Sujet et description

With the growth of digital imaging and 3D printing, color management is becoming the cornerstone of a satisfying visual rendering. However, the wide range of potential interactions between ink, print support, printing technology and light environment makes color management a complex and non-stable task in printing applications. For easier color reproduction, methods based on optical models for spectral predictions have been developed in the recent decades.

In this thesis, a radiativee transfer four-flux model is used for the first time to predict the color of printed materials. Ink layers are colored translucent layers. By considering pigment-based inks are constituted of pigments dispersed inside a binder, we can apply a model of incoherent multiple scattering of light inside the material. Using intrinsic physical and morphological properties of pigments and binder forming the ink, we solve the radiatiive transfer equation with a four-flux model to predict the spectral reflectance and transmittance of ink layers.

The model predictions are compared to spectrophotometric measurements of real printed samples of ink only materrial. We also propose a multiscale solution to predict the visual rendering of inks printed on any support by characterizing separately the spectral properties of the inks and those of the print support.

Finally, with this approach, we introduce a workflow coupling the four-flux model with a surface model for the prediction of any halftone color. Color and spectral deviations between simulated and measured spectra are satisfying while the calibration of the printing system is consiiderably simplified.