ARGIRA v1.0-Experimental: Instrument Characterization Dataset and Software

ARGIRA v1.0-Experimental is the first frozen experimental release of the ARGIRA image-type classification instrument.

The repository contains the original HTML heuristic, a faithful Python implementation used exclusively for characterization, characterization results over a corpus of 162 images, methodological documentation, and reproducibility materials.

The heuristic classifies images into four predefined categories (painting, photograph, AI-generated, and illustration) using deterministic image statistics (irregularity, entropy, edge density, hue entropy, and saturation entropy). No machine learning models or trained classifiers are used.

The characterization corpus contains:

• 108 paintings
• 15 photographs
• 19 synthetic geometric patterns
• 20 real AI-generated images generated with Stable Diffusion and sourced from Lexica

Main characterization results:

• Overall accuracy: 38.9%
• 0% detection of the real AI-generated images included in the corpus
• No high-confidence errors (confidence ≥ 0.75)

This release documents the instrument exactly as evaluated. It is not presented as a state-of-the-art AI image detector or benchmark.

Future improvements will be released as new versions while preserving this frozen baseline for reproducible research.

Repository contents include:

• The frozen original HTML and JavaScript heuristic instrument
• The Python characterization implementation
• Per-image characterization results in CSV and JSON formats, SHA-256 hashes of the corpus, and a feature-space visualization
• The complete characterization report
• The Phase B experimental protocol and Paper 2 theoretical framework
• README.md, RELEASE_NOTES.md, CITATION.cff, and LICENSE.txt

The original corpus images are not redistributed. The repository includes only hashes and metadata to facilitate reproducibility while respecting image licensing.