How GROK Works

GROK Logo

We named this informatics toolkit GROK as an acronym for “general radiation observation kit,” inspired by the word coined by author Robert A. Heinlein in 1961.

Dose Report Identification and Anatomy-Specific Exposure Metric Extraction

GROK locates and retrieves the CT exposure metrics CTDIvol and DLP from a DICOM image archive. Dose report screen captures are retrieved and converted to text using OCR (except for Philips which routinely saves these exposure metrics in private DICOM attributes, obviating the need for OCR).  In addition, it retrieves a single image from each series in the CT encounter in order to obtain image-level DICOM attributes containing information about the scanner, CT protocol, examination, and patient. Finally, anatomy assignment algorithms use the combined dose report screen text and DICOM attributes data to determine the anatomic regions irradiated. All extracted and derived data is written to a relational database (Microsoft SQL server 2005, Microsoft, Seattle, Washington).

Binary Distribution Instructions:

  1. Create a base directory and copy files grok.cfg and GrokBatch.jar to that directory.
  2. Get and install a SQL database if you don't have one already (the freely available PostgreSQL is supported, as is Microsoft SQL Server).
  3. Create a database for GROK to use (default is "grok"). If you are using PostgreSQL, make sure a valid user/password is configured in the database and in grok.cfg. Default authentication for MS SQL Server uses integrated Windows authentication (the Window's user you run GROK as) and you will need to download Microsoft's authentication library sqljdbc_auth.dll and put it in the GROK directory or System path.
  4. Make sure you have Java installed.
  5. Open grok.cfg in a text editor and modify with your directory location, database type and settings, DICOM archive settings, and other runtime settings as needed. (See grok.cfg comments for details.)
  6. Run "java -jar GrokBatch.jar" from the GROK base directory in a command prompt.
  7. Output will be written in your database and (optionally) RDSRs will be created in the RDSR subdirectory.

Source Code:

  1. Peruse, modify, compile, and run the source code at your leisure. GrokBatch is the class intended for batch execution of dose extraction.
  2. N.B. The PixelMed source code included is circa September 2010 and is not up to date with the latest version on David Clunie's page.
  3. Dependency .jar files are in the dependency directory.  They are the same as the PixelMed toolkit ones, plus jdbc drivers for mssqlserver and postgresql databases.

See COPYRIGHT file for the BSD-style open source license and redistribution rules.

Many thanks to David Clunie for developing such a wonderful and open-source DICOM toolkit. 

GROK Related Links


Send Feedback to: Radiology Web Team
This page was last modified on 11/25/2011