Clinical Guide
How to Identify an Unknown Scan Body
When a scan arrives with no implant system documentation, the lab has to figure out which scan body is in the file before production can begin. This guide explains how scan bodies are geometrically distinguished from one another, why the manual identification workflow breaks down, and how automated STL fingerprinting solves the problem.
Why labs receive cases with unknown scan bodies
Dental labs process cases from multiple clinics, each using different intraoral scanners and implant systems. The work order may say "Straumann" — but the scan file itself contains no embedded metadata about which implant platform was actually used. The scan body in the STL is a specific model with specific tolerances. Getting it wrong — designing the restoration to the wrong platform, or ordering the wrong abutment — means the case must be remade.
This problem compounds at scale. With 4,000+ scan body entries across 122+ manufacturers in our reference library — each available in narrow, regular, and wide platform variants — manual identification relies almost entirely on institutional memory and catalog lookups. When that memory is wrong, the remake follows.
What makes scan bodies geometrically distinguishable
Each scan body is a precision-manufactured component, and its geometry encodes the implant system it belongs to. The key distinguishing features are:
Top-cap shape
The anti-rotation geometry at the coronal end — hexagonal, octagonal, trilobular, or proprietary — varies between manufacturers and is the primary connection identifier. Straumann BLT uses a different top-cap profile than Straumann BLX, for example, even though both are Straumann products.
Platform diameter (NP / RP / WP)
Narrow Platform (NP), Regular Platform (RP), and Wide Platform (WP) describe the diameter at the implant-to-abutment interface. Platform diameters differ between narrow and regular platforms by measurable amounts visible directly in the STL geometry — and they determine which prosthetic components are compatible.
Taper angle
The angle at which the body widens from base to shoulder. Conical connections (Morse taper) have a gradual, precise taper. Internal hex connections have a near-vertical wall. The taper ratio is a reliable discriminating feature across platforms.
Height and diameter profile
The full-length cross-sectional diameter profile — measured at regular intervals from base to apex — is unique enough to distinguish even closely related models from the same manufacturer. A Straumann BLT NC and a Neodent GM NF have similar top-cap diameters but very different mid-body profiles.
Apex shape
Whether the top is flat, domed, or has a specific seating feature distinguishes within-brand variants. Some manufacturers use a distinctive apex shape that differs meaningfully from most others — a feature that survives even partial tissue coverage.
Anti-rotation geometry
Features machined into the base that prevent rotation after seating — flat faces, grooves, slots, or indexing features — vary in number, depth, and angular position and provide a strong fingerprint even when other dimensions are similar.
Taken together, these features create a geometric fingerprint that is unique to each scan body model — and that fingerprint is present in the STL file even before anyone looks at it manually.
The manual identification workflow — and where it fails
The current industry standard for identifying an unknown scan body involves several manual steps, each of which introduces error:
- Open the STL in a viewer and visually inspect the top-cap shape and approximate diameter.
- Cross-reference against a catalog — either a printed reference book, a manufacturer's PDF library, or an online catalog — and select the most likely match.
- Open the CAD software, import the suspected scan body from its internal library, and manually align it to the scanned version by placing alignment points.
- Visually verify the fit. If it doesn't match well, repeat with the next candidate.
This workflow has well-documented failure modes:
Why manual identification breaks down
- Visually similar bodies from different manufacturers — particularly in the 3.5–4.0 mm narrow-platform range — are routinely confused. Top-cap shapes that look identical to the eye differ by tenths of a millimeter that determine component compatibility.
- The catalog lookup step depends on the technician knowing where to look. Labs that process cases from 10+ implant systems cannot maintain accurate institutional memory for every platform.
- Manual alignment point selection in exocad and 3Shape requires the user to pre-identify the scan body before importing it. If the pre-selection is wrong, the manual alignment step confirms a false match.
- Partially exposed scan bodies — where soft tissue covers part of the body geometry — make visual inspection and manual alignment significantly less reliable. Accuracy of identification and alignment degrades when a scan body is only partially exposed; this is a recognized limitation across the industry, and manual verification is considered essential even in guided digital workflows.
- There is no checksum. A technician who selects the wrong scan body from the library and aligns it visually gets no objective failure signal — the CAD software proceeds.
What the major CAD platforms require
exocad and 3Shape require the user to manually pre-select the scan body model from an internal library before importing the scan. Alignment of the digital scan body to the patient scan is then performed by manual placement of alignment points, followed by visual verification. The software does not identify the scan body from the geometry — it positions a pre-specified body where the user tells it to.
Shining3D requires the user to manually select the scan body from a database and click "Match" per implant. The matching engine then aligns the selected body to the scan; it does not infer the correct model from the scan geometry.
Every existing workflow requires a person to pick the scan body from a library and/or seed the alignment by hand. ScanBodAI identifies it automatically from the scan geometry — no manual point-picking, no library pre-selection.
How automated STL fingerprinting works
The engine reads the STL geometry directly and measures the scan body across many quantitative features — its profile, taper, platform dimensions, and anti-rotation geometry among them — assembling a geometric fingerprint that is matched against the reference library. No pre-selection is required from the user.
The result is returned with a confidence level: CONFIRMED (strong geometric match with a wide margin over the next candidate), POSSIBLE (plausible match but some ambiguity — verify against the work order), or LOW (no high-confidence match found — the body may not yet be in the library).
No accuracy overclaim: The engine returns a confidence-rated identification, not a guarantee. Partially tissue-covered bodies and previously unindexed models will receive lower confidence scores. The result is a signal to assist the clinician or lab technician — it does not replace verification against the work order and the patient record.
How to identify an unknown scan body with ScanBodAI
Export the STL from your scanner software
In 3Shape, exocad, or any other CAD platform, export the scan body as an STL or OBJ file. The raw mesh is all that is needed — no alignment, no pre-selection.
Upload to ScanBodAI
Drop the file into the ScanBodAI identification tool. Arch scans with multiple scan bodies are supported — the engine segments each body and identifies them independently.
Review the result
The engine returns a ranked list of matches with confidence levels and the geometric features that drove the identification. Compatible prosthetic components — abutments, impression copings, healing caps — are surfaced automatically with catalog numbers.
Verify against the work order
Cross-check the identified manufacturer, platform, and model against what the dentist ordered. A mismatch here — before production starts — prevents the remake.
Try it now — free
Upload an STL and see which scan body it is.
No account required for the demo. No manual point-picking. No library pre-selection.