Scaling perceived saturation.

Area of Science:

• Color science
• Psychophysics
• Perceptual color modeling

Background:

• Accurate prediction of perceived color saturation is crucial for various applications.
• Existing color models require validation with empirical data.
• Understanding observer variability in color perception is essential.

Purpose of the Study:

• To collect replicated perceived saturation data using psychophysical experiments.
• To evaluate the predictive performance of four different color saturation models.
• To assess inter- and intra-observer variability in saturation judgments.

Main Methods:

• Conducted two psychophysical experiments at NCSU and RIT with color-normal observers.
• Used 37 Munsell samples and varied presentation methods (simultaneous vs. single reference).
• Tested four saturation models: Sab*, Suv*, CIECAM02, and Richter/Lübbe, using Standardized Residual Sum of Squares (STRESS).

Main Results:

• CIECAM02 best predicted saturation for NCSU data; Sab* performed best for RIT data.
• Sab*, Richter/Lübbe, and CIECAM02 models showed comparable performance on the combined dataset.
• Inter- and intra-observer variability were low, comparable to color difference studies.

Conclusions:

• Sab*, Richter/Lübbe, and CIECAM02 models offer viable options for predicting perceived color saturation.
• The experimental data provide a reliable basis for future color model development.
• Low observer variability supports the robustness of the collected saturation data.