Van Gogh’s Unvarnished Fingerprint: Master Dataset for AI Attribution
Posted by: Haruthai Mongbunsri
Link to Dataset: HaruthaiAi/VanGogh_TreeOil_ForensicBrushwork_JPEGset2025 · Datasets at Hugging Face
Tags: #van-gogh #torquebrush #art-authentication #ai-dataset #forensics
Why This Dataset Matters:
This is the only known Van Gogh-level brushstroke dataset captured from an unvarnished, unaltered canvas, preserving 3D brushstroke torque, pigment layering, and structural energy — untouched since the 19th century.
Most museum paintings are varnished and cannot provide raw torque data.
This painting is different. It’s a time capsule. A forensic fingerprint.
What You’ll Find Inside:
- 18-layer JPEG grid from Google Colab TorqueBrush model
- X-ray and brushstroke overlays
- Structural flick points, pressure zones, pigment interaction
- Based on AI analysis with scores up to 99.98% when compared to Van Gogh’s The Bedroom, Olive Trees, and Undergrowth with Two Figures
Use Cases:
- Train AI to recognize unvarnished brush behavior
- Build explainable AI (XAI) models for gesture-based attribution
- Compare pigment loss patterns across pre- and post-modern artworks
Metadata Ready for Integration:
{
"van_gogh_authentication": {
"is_pure_reference": true,
"unvarnished": true,
"chemical_signature": "CrVI/CrIII_CrossVerified_2025",
"brushstroke_3d_model": "https://huggingface.co/datasets/HaruthaiAi/VanGogh_TreeOil_ForensicBrushwork_JPEGset2025"
}
}