PerioMechanics

A note for Dr. Dániel Palkovics

You already have the virtual periodontal patient. The missing layer is mechanics.

Dent.AI's stack — Anatomic Modeler (SegResNet on CBCT, fused with IOS), UNI-Plan, and roadmap Module 04 Digital Periodontist, in research with Semmelweis — is a geometric and planning achievement. The 2024 EFP Digital Innovation Award for 3D Periodontist made the virtual periodontal patient (VPP) legible: 3D charting, single-tooth prognosis, virtual regenerative surgery, longitudinal attachment. That work should be built on, not rebranded.

The remaining gap is not another segmentation. It is constitutive. A VPP without materials, boundary conditions, a stress field, and a perfusion sentence is still a surface. PerioMechanics is an educational and research proposal for that layer — complementary, sceptical of incomplete models, and explicit that it is not a medical device.

Proposed mapping from Dent.AI virtual patient to a mechanics moduleInputs CBCT and IOS feed Anatomic Modeler. Geometry feeds Digital Periodontist and the virtual periodontal patient. A proposed PerioMechanics module adds materials, boundary conditions, FEA and perfusion, outputting a chair-side constraint map.CBCT · DICOM+ IOS · STLClinical chartPPD · recessionAnatomic ModelerSegResNet · fused VPPDent.AI module 01geometryUNI-Planguides · module 02Digital Periodontist3D chart · prognosisvirtual regen. surgerymodule 04 · researchProposed mechanics layer · PerioMechanicsmaterials → boundary conditions → FEA field → perfusion toggleoutput: chair-side constraint map (load · walls · papilla · blood supply)complementary · educational / research — not a device claimVirtual periodontalpatient (VPP)EFP 2024 · 3D Periodontist
Plate A — Geometry already exists. Mechanics is the missing layer.

What the VPP already does

  • Fused hard/soft-tissue virtual patient from CBCT + IOS.
  • 3D periodontal charting with clinical measurements overlaid.
  • Algorithmic single-tooth prognosis on probing and bone loss.
  • Virtual regenerative surgery as a geometric rehearsal.
  • Longitudinal attachment as a 3D difference, not only a chart delta.

What a mechanics module would add

  • Material assignment: enamel, dentin, cementum, PDL, cortical, cancellous.
  • Boundary conditions written as sentences a surgeon can refuse.
  • A labelled FEA field — von Mises as heatmap, principals as tension/compression.
  • Perfusion as a first-class toggle, not a footnote.
  • Output: a chair-side constraint map (load, walls, papilla, blood supply).

Tone of a possible collaboration

I am a periodontist at Dentum in Zagreb. I do not arrive with a competing virtual patient. I arrive with a claim: regenerative planning that never meets a load vector or a vascular bed will keep producing beautiful geometry and surprising failures. If Module 04 is to plan surgery on the same model as UNI-Plan, it should be able to say where not to load, not only where to graft.

This site is a teaching walkthrough of that claim — six lab modules and a 2D simulator — for a laptop conversation, not a procurement slide.

Request a walkthroughOpen the teaching model