Dental scan quality control is the crucial first step of checking an intraoral scan for issues before investing any design time into it. The scan is the only physical truth the lab has: every judgment made after it, whether by human or machine, stems from that single file.
What makes a dental scan unusable, and what are the impacts?
It’s not the difficult cases that cause the most trouble; it’s the unreadable ones. Scans fail for everyday reasons: blood or saliva pooling at the finish line, tissue covering a subgingival margin, or missing data in blind spots like a tight contact. Another major issue is an incorrect bite registration, where the upper and lower scans overlap or fail to occlude in the correct relative relationship.
When the input is compromised, the impact is severe. A bad scan makes it impossible to design a crown with a proper margin fit, and you'll inevitably find that the contact or occlusion is completely wrong. Even if a technician carefully fine-tunes the design in the CAD software according to exact parameters, the final restoration will still fail. If the initial input is inaccurate, the output will be inaccurate regardless of the parameters. Wasted bench time, material costs, and frustrating remakes quickly add up.
How do labs currently check for bad scans?
Most labs already attempt a manual version of quality control. A technician opens the file, spins the 3D arch, inspects the prep, and decides if it’s workable before assigning it. This method works, but it simply doesn't scale. In reality, on a busy Monday morning with a full queue, it is usually the first quality check that gets skipped.
How can AI identify bad scans and unreadable data?
There is a fundamental difference between a hard case and a bad scan. A "hard case" means the anatomy is tricky—perhaps a crowded arch or a damaged prep—but the data is still present. A "bad scan" means the data simply isn't there. No amount of AI capability can recover a surface that was never captured.
Therefore, the true value of AI isn't in magically rescuing a bad scan, but in accurately identifying when data is missing or incorrect. Recognition is where this shows up first — the step where the software works out what each surface actually is, and the moment a missing margin or a bad bite first raises a flag. (Dentscape's AI Scan QC, coming soon, puts that check at the front of the queue.)
What does AI scan quality control change for a lab?
To truly scale lab operations, this quality check must be moved to the very front of the workflow. The key is to standardize the process while simultaneously customizing it to fit the specific standards of each individual lab.
By having AI automatically screen a day's scans for problems before any design time is spent on them, labs can flag unusable files immediately. The ultimate benefit is intercepting the error early: an unreadable scan can be sent back to the practice while the patient is still in the chair. This drastically reduces remakes, protects the lab's margins, and eliminates the dreaded phone call a week later to tell the doctor their scan was bad.
Curious about what happens after a scan is cleared? Read more about how AI crown design works, or visit the Dentscape CAD page.