Blur tolerance for luminance and chromatic stimuli

Area of Science:

• Visual perception
• Color vision
• Image processing

Background:

• Human visual system's ability to perceive blur is crucial for image quality assessment.
• Understanding blur tolerance across different color spaces (red-green, yellow-blue, luminance) is essential for accurate visual modeling.

Purpose of the Study:

• To investigate and quantify human blur tolerance across isoluminant red-green, isoluminant yellow-blue, and luminance color directions.
• To compare blur thresholds and sensitivity across these color directions and with existing visual models.

Main Methods:

• Human observers were tested for blur difference thresholds using stimuli modulated along red-green, yellow-blue, and black-white axes.
• Blur thresholds were measured for various reference blurs and stimulus contrasts.
• A computational model (Watt and Morgan, 1984) was used to estimate internal blur of the visual system.

Main Results:

• Blur thresholds for red-green and luminance (black-white) stimuli were similar (0.5 min arc), significantly lower than for yellow-blue stimuli (1.5 min arc).
• Optimal blur thresholds occurred at 1 arc min blur, not sharp edges.
• Weber's law applied to red-green and black-white stimuli for blurs exceeding the optimum.
• Internal blur estimates: 1.2 arc min for black-white and 0.9 arc min for red-green stimuli at 10% contrast.
• Yellow-blue blur tolerance was independent of external blur and not predicted by the model.
• Contrast dependence varied, with yellow-blue showing stronger dependence than red-green and luminance.

Conclusions:

• Visual system's blur tolerance differs significantly across color directions, with red-green and luminance being more tolerant than yellow-blue.
• Findings provide a basis for predicting blur in complex colored images and setting spatial frequency cutoffs for image processing.
• Results offer a baseline for studying visual disorders affecting blur perception, such as amblyopia.