Burned-In Annotation Redaction (OCR) with anonym.plus

Black out patient wording printed onto the pixels, all on device.

Burned-in annotation redaction is the removal of patient wording printed onto the image pixels (DICOM PS3.15). anonym.plus reads it with local OCR. The stamp is then hidden, so the picture no longer shows a name.

When this applies

Older modalities print the name and date right onto the frame. Clearing header tags is not enough. The lettering on the pixels must go too.

How anonym.plus handles it

  1. Open the file in anonym.plus on your own device.
  2. Local OCR reads the lettering baked into the pixels.
  3. The tool boxes each name, date, and ID it finds.
  4. You check each box and add any the eye still sees.
  5. Black out the boxed area, or paint over it.
  6. Save the clean frame. The first copy stays with you.

What you need to provide

PHI entity types detected

Categoryanonym.plus entity typeExample
NamesPERSONR. Castillo → [PATIENT_1]
DatesDATE_TIME07 MAR 2026 → [DATE]
Record IDsMEDICAL_RECORD_NUMBERMRN 661230 → [MRN]
IdentifiersIDAcc 55-7781 → [ACCESSION]
SiteORGANIZATIONHarbor Clinic → [SITE]
AgeAGEAge 64 → [AGE]

Compliance achieved

Anonymize imaging frames offline — see plans & start free →

Limitations & cautions

OCR depends on how clear the stamp is. Faint or warped print can be missed. Always eyeball the result before you share it. A black box covers print but cannot undo a face shown in the frame.

Frequently asked questions

What is a burned-in annotation in imaging?

It is patient print drawn onto the picture itself, not stored in a tag. Old units often stamp a name and date on the frame. Local OCR reads it so you can hide it.

Does the redaction touch the diagnostic area?

Only where you place a box. You review each box first, so the anatomy of interest stays visible while the corner stamps get covered.

Is the OCR step sent anywhere?

No. OCR runs on your own machine. The pixels and the read print never leave the device, so no outside party sees the picture.