Fractal dimensions in perceptual color space: a comparison study using Jackson Pollock's art

Area of Science:

• Art analysis
• Computational aesthetics
• Color science

Background:

• Jackson Pollock's drip paintings exhibit complex visual patterns.
• Analyzing these patterns requires methods sensitive to perceptual color differences.
• Traditional color spaces like red-green-blue (RGB) are not perceptually uniform.

Purpose of the Study:

• To calculate fractal dimensions of color-specific patterns in Pollock's art.
• To compare the effectiveness of L*a*b* color space filtering versus RGB filtering.
• To investigate the impact of color space choice on perceived fractal dimension.

Main Methods:

• Applied a filtering process to color-specific paint patterns in Pollock's works.
• Modeled perceptual response to color differences using the L*a*b* color space.
• Compared results with traditional red-green-blue (RGB) space filtering.

Main Results:

• The L*a*b* color space provides a more consistent definition of "perceptually different" colors.
• RGB filtering underestimates the perceived fractal dimension of lighter patterns.
• The underestimation by RGB filtering was not observed for darker patterns.

Conclusions:

• The choice of color space significantly impacts the analysis of fractal dimensions in art.
• L*a*b* color space is superior for analyzing perceptual complexity in Pollock's paintings.
• Findings relate to aesthetic principles concerning complexity and perception.