From regular text to artistic writing and artworks: Fourier statistics of images with low and high aesthetic appeal

Tamara Melmer, Seyed Ali Amirshahi, Michael Koch, Joachim Denzler, Christoph RediesFrontiers in Human Neuroscience 2013 · DOI

Abstract

The spatial characteristics of letters and their influence on readability have been studied extensively, but global statistical properties relating to aesthetic appeal remain less understood. Natural scenes and many artworks possess a scale-invariant Fourier power spectrum that falls off linearly in log-log plots. This study investigates whether images of text share this property. Analyzing over 1,600 images across Western, East Asian, and Arabic cultures, researchers found that regular text is highly anisotropic and lacks scale-invariance. However, as text moves toward artistic writing (calligraphy) and ornamental art, its statistical properties shift toward the scale-invariant patterns characteristic of fine art and nature, suggesting a universal mathematical signature for aesthetic appeal in human-made patterns.

Introduction

For decades, vision science has focused on letter identification and readability. However, we know very little about the global "look" of text or why certain writing systems are perceived as more beautiful than others. In the natural world, images typically follow a "scale-invariant" pattern where the relationship between coarse structures and fine details remains constant regardless of magnification. This is often represented as a $1/f^2$ power spectrum. This research asks: does text share this "natural" signature, and does artistic writing bridge the gap between functional communication and fine art?

Methods: Analyzing the Visual Signature of Text

The researchers used 2D Fast Fourier Transforms (FFT) to analyze the spatial frequency of 1,611 images. Unlike studies of natural scenes that average all directions, this team analyzed cardinal (horizontal/vertical) and oblique orientations separately to account for the inherent anisotropy of text.

The Fourier power spectrum in the vertical orientation for a text image, illustrating a non-linear "broken" spectrum with a distinct peak.

They measured the slope of the power spectrum at low and high frequencies. A "slope difference" of zero indicates perfect scale-invariance (a straight line in a log-log plot). They also measured anisotropy—the difference in power between horizontal and vertical orientations.

Key Results: The Statistical Bridge to Beauty

The study revealed that regular text is statistically distinct from natural scenes. It features a "broken" spectrum with a shallow slope at low frequencies (low global structure) and a steep slope at high frequencies (high fine detail).

The Fourier power spectrum in the horizontal orientation, highlighting the anisotropy and deviation from scale-invariance found in functional text.

The "Aesthetic Bridge"

As the "artistic claim" of an image increases—from plain text to calligraphy, then to ornaments, and finally to fine art—the slope difference decreases. Artistic writing essentially "smooths out" the statistical spikes of functional text to mimic the scale-invariance of natural scenes.

  • Regular Latin Print: High slope difference (~3.8) and high anisotropy (~1.75).
  • Fine Art: Near-zero slope difference and low anisotropy (~0.5).
  • Calligraphy: Consistently occupied the middle ground, acting as a distinct hybrid of communication and art.

Cross-Cultural Universality

A striking finding was the universality of these patterns. Despite radical differences in script structure, the statistical transition from "functional" to "aesthetic" was mathematically identical across Western, Arabic, and East Asian cultures. This suggests that the human visual system's preference for scale-invariance is a cross-cultural constant.

Conclusion

The authors suggest that regular text must deviate from natural statistics to remain readable; the high-contrast, repetitive lines required for fast information processing are actually "unnatural" and can cause visual strain. In contrast, art prioritizes the "efficient coding" of the human visual system over legibility. Calligraphy and artistic writing serve as a bridge, balancing the functional requirements of language with the mathematical properties of natural beauty.

Summary updated

Abstract

The spatial characteristics of letters and their influence on readability and letter identification have been intensely studied during the last decades. There have been few studies, however, on statistical image properties that reflect more global aspects of text, for example properties that may relate to its aesthetic appeal. It has been shown that natural scenes and a large variety of visual artworks possess a scale-invariant Fourier power spectrum that falls off linearly with increasing frequency in log-log plots. We asked whether images of text share this property. As expected, the Fourier spectrum of images of regular typed or handwritten text is highly anisotropic, i.e., the spectral image properties in vertical, horizontal, and oblique orientations differ. Moreover, the spatial frequency spectra of text images are not scale-invariant in any direction. The decline is shallower in the low-frequency part of the spectrum for text than for aesthetic artworks, whereas, in the high-frequency part, it is steeper. These results indicate that, in general, images of regular text contain less global structure (low spatial frequencies) relative to fine detail (high spatial frequencies) than images of aesthetics artworks. Moreover, we studied images of text with artistic claim (ornate print and calligraphy) and ornamental art. For some measures, these images assume average values intermediate between regular text and aesthetic artworks. Finally, to answer the question of whether the statistical properties measured by us are universal amongst humans or are subject to intercultural differences, we compared images from three different cultural backgrounds (Western, East Asian, and Arabic). Results for different categories (regular text, aesthetic writing, ornamental art, and fine art) were similar across cultures.
BibTeX
@inproceedings{melmer2013from,
  author = {Tamara Melmer and Seyed Ali Amirshahi and Michael Koch and Joachim Denzler and Christoph Redies},
  title = {{From regular text to artistic writing and artworks: Fourier statistics of images with low and high aesthetic appeal}},
  booktitle = {Frontiers in Human Neuroscience},
  year = {2013},
  doi = {10.3389/fnhum.2013.00106},
}
Figure