pydicom / dicom-cleaner

detection of burned in pixels using OCR (under development)

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Dicom Cleaner

This repository provides tools intended for use to scrape personal health information (PHI) that is represented as text from image headers and pixels. Each subfolder here corresponds to a different kind of cleaner, and a docker container.

  • ocr "optical character recognititon" is an image that runs text (letter detection) on a demo image. You can either detect (just find and report) or clean the data (and save cleaned png images).
  • header uses the deid python module to target known coordinates based on manufacturer and machine types to flag images. The coordinates are then cleaned (covered with a black box).

Both containers are available on Docker Hub.

About

detection of burned in pixels using OCR (under development)


Languages

Language:Limbo 59.6%Language:Objective-C 40.0%Language:Python 0.4%Language:Dockerfile 0.0%Language:Shell 0.0%