Data De-identification

Segmed Incognito
Our fully automated pipeline identifies and redacts protected health information (PHI) from DICOMs and medical text reports.

DICOMs, de-identified


At a high level, Incognito uses OCR to detect text (including burnt text & handwritten notes) on DICOM images.


Automatically black-out identifying information from DICOM images, ensuring the image is safe to use.

Text data, de-identified

We de-identify medical text reports using our proprietary language transformer model.

Direct and indirect patient and institution identifiers are redacted to ensure effective de-identification.
99.97% success rate of text de-id under HIPAA Safe Harbor standards
Get in touch

Learn more:

How is it made available?

Our technology is configurable - we offer de-identification as a hosted service, and as standalone SaaS.

Who is Incognito for? 

Enterprise orgs, startups, research groups, universities, hospitals and more!
Anyone who needs to adhere to regulations and de-identify their data.

How can I get started?

Get in touch with our team to schedule a demo. You can also take a look at and test our LLM-based de-id playground.

Accelerate your algorithm development pipeline

Curate your training & validation datasets on Segmed Insight today!