Premium
Artificial neural networks for optimising camera‐based colour measurements of prints enhanced with pearlescent pigments
Author(s) -
Tomić Ivana,
Dedijer Sandra,
Novaković Dragoljub,
Jurič Ivana
Publication year - 2018
Publication title -
coloration technology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.297
H-Index - 49
eISSN - 1478-4408
pISSN - 1472-3581
DOI - 10.1111/cote.12346
Subject(s) - artificial intelligence , digital camera , computer science , computer vision , flexibility (engineering) , artificial neural network , viewing angle , single camera , optics , mathematics , physics , statistics , liquid crystal display , operating system
Pearlescent pigments are widely used in printing due to their optical, chemical and physical properties. To analyse the effects of goniochromism they produce, the colorimetric characterisation of materials printed with pearlescent pigments requires multi‐angular measurements. In this study, the colours of prints enhanced with pearlescent pigments were measured by means of a digital camera, relying on the empirical camera characterisation method. Since this method is time‐consuming, it was altered to enable estimates of colorimetric values for different geometries to be measured on the basis of images captured at one viewing angle. This approach was based on the use of artificial neural networks which were shown to provide sufficient flexibility for the given task. The results indicate that the images obtained at the viewing angle of 45° aspecular (measuring geometry 45°/asp 45°) accurately estimate CIEL ab values for all of the tested measuring geometries. The proposed method is therefore not only time‐efficient but also reduces the associated errors due to the camera's movement, and enables the estimation of colorimetric values for those viewing angles inaccessible by camera.