Fractal contours: Order, chaos, and art.

Area of Science:

• Art analysis
• Computational geometry
• Image processing

Background:

• Traditional methods use single parameters like fractal dimension to characterize artwork complexity.
• These methods often rely on fitting data to nonlinear scaling plots, leading to inaccuracies.
• Existing fractal analysis techniques are sensitive to methodological choices and image orientation.

Purpose of the Study:

• To develop a more robust and discriminating method for analyzing the structural and topological properties of artworks.
• To extract relevant information from scaling plots without relying on fitting routines.
• To characterize images using a continuous function, the fractal contour, instead of a single exponent.

Main Methods:

• Averaging over all possible grid locations at each scale to ensure scaling plot independence from grid choice and image orientation.
• Calculating derivatives of the scaling plots to generate a continuous fractal contour for each image.
• Testing the method on synthetic images, algorithmic fractals, and artworks by modern masters.

Main Results:

• The proposed method yields robust scaling plots, independent of grid orientation.
• The fractal contour provides a more detailed characterization than a single fractal exponent.
• The method successfully distinguishes between different types of images, including complex artworks.

Conclusions:

• The fractal contour method offers a more discriminating approach to analyzing artwork complexity and topology.
• This technique overcomes limitations of traditional single-parameter fractal analysis.
• The method has potential applications in art classification and understanding artistic style.