Are you an EPFL student looking for a semester project?
Work with us on data science and visualisation projects, and deploy your project as an app on top of Graph Search.
The Yule-Nielsen modified spectral Neugebauer model (YNSN) enables predicting reflectance spectra from ink surface coverages of halftones. In order to provide an improved prediction accuracy, this model is enhanced with an ink spreading model accounting for ink spreading in all superposition conditions (IS-YNSN). As any other spectral reflection prediction model, the IS-YNSN model is conceived to predict the reflection spectra of uniform patches. Instead of uniform patches, we investigate if tiles located within color images can be accurately predicted and how they can be used to facilitate the calibration of the ink spreading model. In the present contribution, we first detail an algorithm to automatically select image tiles as uniform as possible from color images by relying on the CMY or CMYK pixel values of these color tiles. We show that if these image tiles are uniform enough, they can be accurately predicted by the IS-YNSN model. The selection algorithm incorporates additional constraints and is verified on 6 different color images. We finally demonstrate that the ink spreading model can be calibrated with as few as 5 to 10 image tiles.
Sabine Süsstrunk, Majed El Helou, Frederike Dümbgen, Natalija Gucevska
,