Home // CONTENT 2023, The Fifteenth International Conference on Creative Content Technologies // View article

Color Manipulation in Images by Using a Modified HSV Algorithm

Authors:
Samuel Kosolapov

Keywords: Image Processing; HSV; color manipulations; reliable color recognition.

Abstract:
There are several situations in which the HSV (Hue, Saturation, Value) algorithm is preferred to execute color manipulations in a specific image. Unfortunately, whereas Value and Saturation can be calculated in any situation, Hue cannot be calculated when the specific pixel is gray. Even more, for dark regions, nearly gray regions, and for overbleached regions, the calculation of the Hue is non-reliable. Inherent to digital image noise makes the calculations of Hue in the above situations even more problematic. In an attempt to provide more control during color manipulations, an extended structure “sHSVF” was defined, in which F is a "Validity Flag". Additionally, the structure "sValidityParameters" was defined. Values of the fields in this structure make it possible to classify the specific pixel as "IS_VALID", “HAS_ZERO”, "IS_TOO_GRAY", IS_OVERBLEACHED", "IS_TOO_DARK", "IS_TOO_LIGHT", and properly set the "Validity Flag" for each pixel of the image. This flag may be instrumental in color recognition and in reliably modifying the color of the pixels in accordance with the selected rules. By selecting values of the “sValidityParameters”, the user of the algorithm can specify situations when the color of the specific pixel is set to the pre-defined value, marking problematic situations. Provided examples demonstrate that this approach can be used for reliable color recognition and advanced color manipulations for synthetic and real-life images.

Pages: 1 to 6

Copyright: Copyright (c) IARIA, 2023

Publication date: June 26, 2023

Published in: conference

ISSN: 2308-4162

ISBN: 978-1-68558-048-3

Location: Nice, France

Dates: from June 26, 2023 to June 30, 2023