FeaturesAI Scan Quality Control
Three engines, one verdict

AI Scan Quality Control

Evaluate dental STL scans using AI-powered analysis before they enter production. Three evaluation engines — native AI, Movix, and BiteFinder — provide comprehensive defect detection, mesh validation, hole analysis, occlusal evaluation, and region-of-interest painting with interactive 3D brush tools.

AI Engines
3 engines
3D Viewer
Interactive
ROI Painting
Face-level
Training Mode
AI vs. human

Native AI Evaluation

Leveonn's built-in AI engine scores scans across multiple configurable criteria. Set weights per criterion, define pass/fail thresholds, and let the AI deliver consistent, objective quality verdicts on every scan.

Multi-criterion scoring with configurable weights per evaluation category

Overall pass/fail verdict with detailed per-criterion breakdown

Configurable evaluation templates for different scan types

Historical score tracking and trend analysis per patient

Batch evaluation mode for processing multiple scans simultaneously

Movix Integration

Deep integration with Movix's dental scan analysis platform. Create cases, upload STLs, and run a full suite of validation tasks — mesh validation, hole detection, defect analysis, and hyperocclusion evaluation — all from within Leveonn.

Automated case creation and STL upload via presigned URLs

Validation task: mesh integrity and surface quality checks

Hole detection task: identifies gaps and missing geometry

Defect analysis task: locates surface defects and anomalies

Hyperocclusion task: evaluates occlusal contact accuracy

Run-all mode for batch execution of every task type

3D viewer link generation for detailed visual review in Movix

BiteFinder Integration

Automated bite analysis powered by BiteFinder's specialized engine. Upload scan files, configure analysis parameters, and retrieve results — all with automatic status polling and completion detection.

Case creation with configurable file types and feature parameters

Presigned URL upload proxy for secure, large file transfers

Automatic status polling with completion detection

Analysis results retrieval with downloadable output files

Per-case job configuration JSON for customized analysis parameters

Local case tracking in Firestore for real-time UI state

3D STL Viewer & ROI Painting

A full Three.js-powered 3D viewer built directly into the platform. Rotate, zoom, and inspect scans in real-time. Paint regions of interest on the mesh surface using a brush tool that selects individual faces for annotation.

Real-time 3D rendering with orbit controls, zoom, and pan

Brush tool for painting regions of interest directly on mesh faces

Face-level selection stored in Firestore subcollection for scalability

Multiple ROI colors for categorizing different region types

Screenshot capture of annotated 3D views for documentation

Smooth performance with large STL files via optimized geometry handling

Batch Upload & Patient Tracking

Process multiple scans at once with batch upload workflows. Track evaluations at the patient level, maintain history across visits, and build a comprehensive quality record for every case.

Drag-and-drop batch upload for multiple STL files

Patient-level scan organization with visit history

Evaluation history with score trends over time

Cloud Storage integration for secure scan archival

Cross-engine comparison showing results from all three evaluation sources

Technician Training Mode

Train your QC team by comparing human evaluations against AI verdicts. Technicians perform manual assessments, then see how their judgment aligns with the AI — building consistency and confidence across your team.

Side-by-side comparison of human vs. AI evaluation scores

Per-technician performance tracking and improvement metrics

Configurable training exercises with selected scan sets

Feedback loop that helps calibrate both human and AI judgments

Training history with progress visualization over time

How it works

The typical workflow from start to finish.

1
Upload STLs
Drag and drop scan files or sync from IOS intake
2
Select Engines
Choose which AI engines to run: native, Movix, BiteFinder
3
Run Evaluation
AI engines analyze scans for defects, holes, and quality issues
4
Review in 3D
Inspect results in the interactive 3D viewer with annotations
5
Paint ROIs
Use brush tools to mark regions of interest on the mesh
6
Approve or Reject
Make the final QC decision with full evaluation context

See AI Scan Quality Control in action

Every feature is live and production-ready. Let us walk you through how it fits into your lab.