Multi-Channel Fractal Analysis: Unveiling Stylistic Complexity in Digital Art

Title: Multi-Channel Fractal Analysis: Unveiling Stylistic Complexity in Digital Art (Source Code and Experimental Data)

Overview:
This dataset contains the source code, precomputed results, and analytical tools developed for the quantitative characterization of artistic styles through multi-channel fractal analysis. The research addresses the limitations of grayscale-based complexity measures by implementing a multi-channel fractal dimension (FD-RGB) framework based on the RGB Differential Box-Counting algorithm.

Content:
The repository is organized to ensure full reproducibility of the study "Multi-Channel Fractal Analysis: Unveiling Stylistic Complexity in Digital Art" (submitted to The Visual Computer). It includes:
1. Source Code: A high-performance implementation of the RGB Differential Box-Counting algorithm using CUDA/C++ for processing large-scale image datasets, along with MATLAB scripts for image preprocessing (cropping, format conversion) and statistical analysis.
2. Experimental Results: Comprehensive CSV files containing the computed metrics (FD-Gray, FD-RGB, Chromatic Complexity Gain ΔFD, and Shannon Entropy) for the ArtBench-10 dataset, which comprises 60,000 paintings across ten distinct artistic styles.
3. Statistical Analysis: Tools for evaluating stylistic divergence, classification performance, and the mapping of computational signatures to canonical pictorial techniques.

Purpose:
This resource is intended for researchers in computational aesthetics, computer vision, and digital humanities. By providing both the raw algorithmic tools and the precomputed complexity metrics, it facilitates the objective study of visual complexity in digital heritage and the development of new stylistic discrimination models.

Citation:
This dataset is an integral part of the following research. If you use these resources, please cite:
Ruiz de Miras, J., and Martín, D. "Multi-Channel Fractal Analysis: Unveiling Stylistic Complexity in Digital Art", The Visual Computer, 2026.