Gaussian-Based Hue Descriptors

Area of Science:

• Color Science
• Computer Vision
• Human Perception

Background:

• Accurate hue representation is crucial for color modeling and computer vision.
• Existing models like CIECAM02 have limitations in stability across different lighting conditions.

Purpose of the Study:

• To explore a novel hue descriptor for improved color perception modeling and computer vision.
• To develop a descriptor that is robust and stable under various illuminants.

Main Methods:

• Utilized a Gaussian-like function termed a wraparound Gaussian to define the hue descriptor.
• Correlated the new descriptor with established color atlases (Munsell, Natural Color System) and a color thesaurus (Moroney's).
• Assessed the descriptor's stability across a range of illuminants.

Main Results:

• The new hue descriptor demonstrates strong correlation with established hue designators and names.
• The wraparound Gaussian descriptor exhibits significantly greater stability under varying illuminants compared to CIECAM02.
• This model addresses limitations found in previous Gaussian-based hue modeling approaches.

Conclusions:

• The wraparound Gaussian hue descriptor offers a robust and accurate alternative for color science and computer vision.
• Its enhanced stability makes it particularly suitable for applications requiring consistent color representation under diverse lighting.
• This work advances hue modeling by providing a more reliable and perceptually relevant descriptor.