Open work, where it helps the field.
We share what we can without compromising the closed-source product. Open utility code, synthetic datasets, methods writing, and partnerships with academic dental research groups.
aidcam-tools, utilities for AI dental CAD
A small Python package of utilities for working with dental 3D files and evaluating AI-CAD methods. MIT-licensed. No model code, no AID·CAM proprietary architecture, just the plumbing every dental ML project ends up rewriting.
-
Mesh-to-point-cloud sampling
aidcam-mesh-to-pc CLI and Python API. Loads STL/OBJ/PLY via trimesh, samples N points uniformly from the surface, writes to NPZ. Deterministic seed for reproducibility.
-
Distance metrics for dental geometry
Chamfer L1, Chamfer L2, symmetric Hausdorff. Numpy + SciPy under the hood, no PyTorch dependency. Suitable for benchmarking model outputs against ground truth in millimetres.
-
FDI tooth-numbering helpers
is_molar(16), opposite(11), contralateral(35). Five-line utilities every dental ML project needs and none of them ship.
Synthetic prep + crown benchmark, coming soon to Zenodo.
We are preparing a public release of ~100–200 paired synthetic preparations and crowns, generated by shrink-along-normals from Teeth3DS+ natural teeth. Each case is labeled with FDI tooth number, material annotations, and margin-line ground truth. License: CC-BY 4.0. The corpus is explicitly not a substitute for real clinical data, the distributional gap on subgingival margins and chamfered preparations remains real, but it provides a reproducibly evaluable floor for method comparison.
We work with academic dental research groups.
If you are at a university dental school, a graduate prosthodontics program, or a research lab with archived clinical case data, we'd like to talk. Our framing is reciprocity, not extraction: in exchange for access to clinical data under whatever agreement your institution requires, we share back trained model weights, synthetic dataset expansions, co-authorship on methods papers, and engineering labor on data-pipeline work.
We're building partnerships with academic dental-research groups and clinical-data holders in North America and beyond.