Premium
Categorizing color shifts due to tinted glazing via dominant colors of the scene
Author(s) -
Raza Aiman,
Jost Sophie,
Dubail Marie,
Dumortier Dominique
Publication year - 2021
Publication title -
color research and application
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.393
H-Index - 62
eISSN - 1520-6378
pISSN - 0361-2317
DOI - 10.1002/col.22629
Subject(s) - glazing , computer vision , artificial intelligence , rgb color model , computer graphics (images) , computer science , hue , rendering (computer graphics) , color difference , color model , primary color , color balance , color space , color image , image processing , image (mathematics) , engineering , enhanced data rates for gsm evolution , civil engineering
Tinted films and glazing application on windows reduce glare but can sometime induce color shifts, which needs to be characterized by fast and reliable means. We propose a color graphic that uses the dominant colors of the image seen through the tinted film and compares it to the dominant colors of the original unfiltered image. An image segmentation algorithm applied on a Gaussian blurred image segments the whole scene into six clusters. The median RGB values of the six clusters provide the six dominant colors, which were validated through the means of a psycho‐visual experiment. This difference in the dominant colors of the original and the tinted scene is visualized through a color vector connecting the two colors on the CIELAB's a * b * color plane. This color graphic was directly applied to a glazing application and the predictions of the color graphic visually matched the actual rendition of the glazing. This color graphic provides a clear and concise information about the color shifts that can be directly applied in building and lighting industry to study final renditions of lighting and interior glazing applications. The graphic is especially useful for highly chromatic glazing, which cannot always be characterized by traditional color rendering, because they are too distant from the Planckian/black body curve.