A Computational Model for Color Constancy by Separating Reflectance and Illuminant Edges within a Scene

Area of Science:

• Computer Vision
• Computational Neuroscience
• Image Processing

Background:

• Color constancy is crucial for visual perception, enabling stable color recognition under varying illumination.
• Existing color constancy models struggle with shadowed scenes due to abrupt illuminant changes, often confusing them with reflectance variations.

Purpose of the Study:

• To develop a computational model for color constancy that effectively differentiates between shadows and reflectance changes.
• To improve the accuracy of color constancy algorithms in complex, shadowed environments.

Main Methods:

• A novel computational model employing two modules to separate reflectance and illuminant edges.
• Integration of line processes within each module to interpret edge origins based on shadow properties (luminance changes at boundaries).
• Simulation-based evaluation of the model's performance in edge detection and illuminant color recovery.

Main Results:

• The proposed model accurately distinguishes between reflectance and illuminant edges in shadowed scenes.
• Successful recovery of illuminant colors and surface reflectances.
• Effective removal of illuminant effects from input scenes, demonstrating robust color constancy.

Conclusions:

• Separating reflectance and illuminant changes is a viable strategy for achieving color constancy in shadowed conditions.
• The developed model offers a significant advancement in handling complex lighting scenarios for computer vision and perception research.