ABANICCO: A New Color Space for Multi-Label Pixel Classification and Color Analysis.

Area of Science:

• Computer Vision
• Color Science
• Image Processing

Background:

• Accurate pixel classification and segmentation are crucial for color image analysis in computer vision.
• Challenges exist in aligning human color perception, linguistic terms, and digital color representations.
• Existing methods struggle to bridge the gap between subjective human color experience and objective digital data.

Purpose of the Study:

• To introduce ABANICCO (AB ANgular Illustrative Classification of COlor), a novel unsupervised method for automatic pixel color classification and naming.
• To develop a robust strategy for color naming that is unbiased and grounded in color theory and statistics.
• To provide a standardized and understandable alternative for color analysis recognizable by both humans and machines.

Main Methods:

• Combines geometric analysis, color theory, fuzzy color theory, and multi-label systems.
• Classifies pixels into 12 conventional color categories.
• Evaluated against the ISCC-NBS color system and state-of-the-art image segmentation methods.

Main Results:

• Demonstrated accuracy in color detection, classification, and naming.
• Showcased reliable performance in image segmentation tasks.
• Provided empirical evidence of ABANICCO's effectiveness and robustness.

Conclusions:

• ABANICCO offers a standardized, reliable, and understandable approach to color naming.
• The model serves as a foundational tool for diverse computer vision applications.
• Successfully addresses challenges in region characterization, histopathology, fire detection, and more.