CBCT Segmentation for Dental Laboratories
GPU-accelerated anatomical segmentation with automatic landmark detection
49
Structures
~7m
Processing
STL
Output Format
REST
API
Automatic Landmark Detection
Fiducial Points for CBCT-to-IOS Registration
Anatomical landmarks are automatically identified during segmentation, providing registration points for aligning CBCT data with intraoral scans. Eliminates manual landmark placement and reduces alignment errors in the merge workflow.
Detected landmarks include:
- Inferior alveolar canal entry/exit points
- Mental foramen locations
- Tooth cusps and root apices
- Bone ridge contours
- Maxillary sinus floor boundaries
Input Specifications
| Parameter | Specification |
|---|---|
| File Formats | .dcm, .zip, .tar.gz, .7z, .rar |
| Max File Size | 2 GB |
| DICOM Compatibility | All major CBCT manufacturers (Planmeca, Carestream, Sirona, Vatech, i-CAT, Morita, KaVo, NewTom) |
| Optimal Voxel Size | ≤200μm for surgical planning, ≤300μm for diagnostic |
| Field of View | Full arch, half arch, or quadrant scans supported |
Output Specifications
| Parameter | Specification |
|---|---|
| File Format | .stl (binary) |
| Coordinate System | Patient coordinate system (RAS), centered at scan origin |
| Units | Millimeters (mm) |
| Mesh Quality | Watertight, manifold meshes suitable for CAD/CAM |
| Delivery | Individual STL per structure + JSON manifest with metadata |
Segmented Structures (49 Total)
Bone Structures
Upper Jawbone (Maxilla)
Lower Jawbone (Mandible)
Critical Structures
Left Inferior Alveolar Canal
Right Inferior Alveolar Canal
Maxillary Sinus (L/R)
Upper Teeth (FDI 11-18, 21-28)
Central Incisors
Lateral Incisors
Canines
First Premolars
Second Premolars
First Molars
Second Molars
Third Molars
Lower Teeth (FDI 31-38, 41-48)
Central Incisors
Lateral Incisors
Canines
First Premolars
Second Premolars
First Molars
Second Molars
Third Molars
Additional Structures
Pulp Chambers
Existing Crowns
Existing Implants
API Integration
RESTful API with webhook notifications for async processing.
Upload Endpoint
POST https://imaging.alveaos.com/api/imaging/v2/upload # Headers Authorization: Bearer {api_key} Content-Type: multipart/form-data # Form Data file: {dicom_archive} patient_name: "John Smith" patient_id: "PT-12345" # optional
Response (JSON)
{
"case_id": "b756404c-2a89-4501-a669-79e9718e684c",
"status": "processing",
"estimated_completion": "2026-02-04T15:30:00Z",
"webhook_url": "https://your-lab.com/webhook"
}
Manifest (on completion)
{
"case_id": "b756404c-2a89-4501-a669-79e9718e684c",
"status": "completed",
"processing_time_ms": 412000,
"quality_tier": 2,
"voxel_size_um": 150,
"structures": [
{
"type": "lower_jawbone",
"confidence": 0.97,
"volume_mm3": 45230.5,
"download_url": "https://..."
},
// ... 48 more structures
],
"landmarks": [
{
"type": "mental_foramen_left",
"position": [-23.5, 12.1, -8.3]
},
// ... more landmarks
]
}
Lab Workflow Integration
| Step | Action | Details |
|---|---|---|
| 1 | Receive DICOM from clinician | Via portal upload, email, or direct transfer |
| 2 | Submit to Alvea API | Automated via integration or manual upload at imaging-demo |
| 3 | Receive STL package | ~7 minutes, webhook notification on completion |
| 4 | Import to ExoCAD/3Shape | Individual STLs load directly into CAD software |
| 5 | Merge with IOS scan | Use detected landmarks for registration alignment |