Finds identifiers left in burnt-in pixels and English free text after ingest anonymization. Labels: NAME, LOCATION, ID, PHONE, DATE, and EMAIL.
Download installs OCR and NER model trees together. PHI text detection uses an English NER model for all UI languages (Roman/Latin script).
Use Series View for burnt-in text (Blackout / Inpaint). Use Report Text Review for structured reports and PDFs (Blank or Synthetic hide-in-plain-sight). Capture stores removed PHI entity text only — never full report transcripts.
Ingest never runs NER/OCR. Tag keep/remove is controlled by the anonymizer script and the project de-identification options (Basic or Retain).