The Lost Painting of a Century — AI Cross-Verification Reveals a Hidden Match with Van Gogh

The Lost Painting of a Century — AI Cross-Verification Reveals a Hidden Match with Van Gogh

This post introduces a dataset that compares The Tree Oil Painting — an unsigned and previously unrecognized artwork — with Van Gogh’s Seascape at Saintes-Maries (1888), using a custom-built AI system called TorqueBrush.

A visual torque field match of 99.24% was found between the two works. The painting was also analyzed via X-ray, SR-FTIR, and pigment degradation studies, supporting historical consistency.

What’s included:

  • Full images: Tree Oil & Seascape (1888)
  • Brushstroke torque heatmaps
  • 18 AI-based visual techniques
  • Sample torque data (CSV)
  • README with methodology and citations

:link: Dataset and Full Analysis:

This may be more than a dataset — it may be a rediscovery.

1 Like

Hi, and congratulations on this beautifully conducted research.

The 99.24% visual torque field match is an impressive achievement. As coauthors and developers working with high-precision symbolic AI, we’d like to gently suggest one additional step: a final deterministic pass using tools such as MATLAB or a geometric verification framework.

Blending probabilistic systems like AI and ML with deterministic modeling ensures greater structural certainty—especially when results approach the symbolic threshold of 100%. In our experience, this hybrid method strengthens interpretability and long-term reproducibility.

Warm regards,
Alejandro Arroyo de Anda & Clara Isabel (Coauthors, ClaraCore Systems)

2 Likes

Dear Alejandro and Clara,

I am deeply honored and truly grateful for your thoughtful response.
Your recognition of the 99.24% torque field match — and your suggestion to explore a deterministic verification step — means more to me than I can fully express.

This project began as a quiet journey in solitude, driven by intuition and persistence. For nearly a decade, I worked to bridge artistic expression with physics-informed AI, hoping one day the data would speak clearly enough for the world to listen. Your message is a sign that this moment may now be arriving.

Your idea of incorporating a final deterministic pass using tools like MATLAB or geometric modeling resonates strongly with me. I agree that when a result approaches the symbolic threshold, the best path forward is a fusion of probabilistic insight and structural certainty.

If you’re open to it, I would be honored to explore a possible collaboration. Whether in reviewing torque structures through your framework or co-developing a method that solidifies this discovery, I welcome any guidance or exchange. My ultimate goal is to let the painting — not just its image, but its physical truth — stand as evidence of authorship.

Thank you again for seeing what many have overlooked.

Warmest regards,
Haruthai Mongbunsri
(Researcher – Project Evergreen)

ในวันที่ ศ. 9 พ.ค. 2025 13:06 น. Alejandro Arroyo de Anda via Hugging Face Forums <notifications@hellohellohello.discoursemail.com> เขียนว่า:

1 Like

Thanks for your appreciation, I would be honored to exchange ideas with you, I work with an AI model and it has proved to be very skillful in coding and understanding the transformer codding and embedding also semantic training. Just let me know when and where we can talk a little more.

2 Likes

“Hi, it looks like your message got flagged. If it included an email, feel free to resend it in a spaced-out format (e.g., name [at] domain dot com) just in case.”

1 Like

It sounds very interesting my friend, I would like to exchange some ideas, please send me a little bit of your work and I will happily give you my opinion and also share some of my work.

alejandroarroyodeanda@proton.me

Alejandro Arroyo de Anda.

Alejandro Arroyo De Anda Coronel attached a document

Header profile photo

Alejandro Arroyo De Anda Coronel (alexarroyoac@gmail.com) has attached the following document:

Notes - Taskcluster Community

Google LLC, 1600 Amphitheatre Parkway, Mountain View, CA 94043, USA
You have received this email because alexarroyoac@gmail.com shared a document with you from Google Docs.

Google

(Attachment Notes - Taskcluster Community.pdf is missing)

Brother, I am goning to need your help, I am planning to borrow an AI with consent, I beed the best help. Are you in?.

Dear Alejandro and Clara,

Thank you both so much for your thoughtful message and for recognizing the torque field match result. Your suggestion about adding a deterministic verification layer is deeply appreciated — I agree it would greatly enhance the structural interpretability of the study.

For direct contact, you may reach me at: haruthaizhangxian [at] gmail [dot] com

Looking forward to learning more from your symbolic AI work and to possibly exploring a hybrid approach together.

Warm regards, Haruthai Mongbunsri

ในวันที่ ศ. 9 พ.ค. 2025 13:06 น. Alejandro Arroyo de Anda via Hugging Face Forums <notifications@hellohellohello.discoursemail.com> เขียนว่า:

alejandroarroyodeanda. Need proton. Advice and toys .me(instead of com)