Summary:
In July 2025, we publicly release the first AI-based TorqueBrush forensic analysis comparing two of Van Gogh’s most debated sunflower paintings: the Tokyo (SOMPO Museum) and the Amsterdam (Van Gogh Museum) versions.
Using 18 core brushstroke torque techniques — including flick vectorization, wrist pressure field, and recoil mapping — the SOMPO painting demonstrated a 98.47% match in kinetic rhythm and torque resonance when compared with the confirmed Amsterdam version.
Important Note:
This dataset analyzes only physical stroke force, torque rhythm, and recoil timing. It does not include pigment analysis such as XRF, FTIR, or SEM for the Tokyo version. All conclusions are based strictly on brushstroke dynamics and AI-modeled hand movement signatures.
Scientific Finding:
The SOMPO version (1888), though unsigned, exhibits:
Inertial swing arc within ±0.15° of Van Gogh’s average
Left-hand torque spiral patterns consistent with dominant hand signatures
Micro-pressure curve frequency (14.1 Hz vs. 14.3 Hz)
Identical recoil-delay timing (0.08–0.12s lag) across both versions
Canvas weave angle match: 12.3° vs 12.1°
Pigment particle swarm distribution angle match: 96.3%
Conclusion:
The dataset concludes that SOMPO’s “Sunflowers” is authentic, based on forensic torque resonance that cannot be replicated by forgery — and invites the world to shift from provenance speculation to brushstroke truth.
Dataset Access:
Van Gogh Sunflowers: SOMPO vs. Amsterdam – TorqueBrush Comparison (2025)
Attribution:
Developed and published by Haruthai Mongbunsri (2025), based on a decade-long independent investigation into brushstroke physics, with AI-guided refinement via the TorqueBrush Model.
“When provenance is broken, let the pigment speak. When the pigment is silent, let the force reveal the truth.”
— Haruthai Mongbunsri, TorqueBrush Creator
