DICOM : import and query of medical images

Document created by resplin Employee on Jun 6, 2015
Version 1Show Document
  • View in full screen mode

Obsolete Pages{{Obsolete}}

The official documentation is at: http://docs.alfresco.com



Community Contribution Ideas


Project Description


Add Functionality to import and search for(series and sequences of) medical images (DICOM -conformant files containing CT, PET or MRI images). All DICOM metadata attributes like image modality type, equipment number, acquisition parameters, image resolution and clinical trial study data should be parsed and stored as Lucene indexes. Also create, store and show Thumbnails of the DICOM files. Build an advanced search form for querying all the metadata. Interface with DICOM visualisation software.

Contact Address: Frans Cornelissen, Technology Office, Pharma R&D IM, J&J
tel 32-14-603327 email fcorneli@prdbe.jnj.com


PROJECT REQUIREMENTS


Project Requirements & Objectives


  • Upload a single or a series (sequence) of DICOM images into Alfresco
  • During upload, all DICOM metadata is parsed and indexed : Create Rules and Actions for this
  • During upload, a thumbnail image (jpeg2000) is constructed and linked bidirectionally with the original full-size image (build onto the Alfresco-Thumbnails project in the Forge see http://forge.alfresco.com/projects/thumbnails/ )
  • add a jpeg2000 plugin to IE/FireFOx or make Alfresco handle this
  • A special 'thumnail view' to view the thumbnails of DICOM or other images (adapted icon view?)
  • It shall be possible to choose the size for the thumbnails during creation and display (using jpeg2000 thumbnails, they can be displayed at any size)
  •   Create Rules and Actions for automatic upload/import of complete folders or acp files (in the background) when placed in an 'IMPORT' space; via CIFS, WebDav and Web browser.
  • When clicking on a thumbnail (or original image icon/filename), the image or series should be opened within a new browser window with a DICOM visualisation applet (RadScaper, RemotEye, other OS applets)
  • When editing or annotating the DICOM images within the viewer applet, or creating voice or text reports, these additions or changes should be written back into the repository (new file version?)
  • A (tabbed?)query form made specifically to rapidly enter DICOM metadata to search for:
    • for all attributes in all DICOM metadata groups (from the official DICOM Data Dictionary)
    • it shall be possible to select 'fixed choices' (enumerated possibilities) from a drop-down list

PERFORMANCE REQUIREMENTS


  • Performance benchmark within a full Alfresco design (ie images stored & managed by Alfresco) – with 10M images into the Alfresco repository, splitted over 10 000 folders, and representing approximately 1 up to 2 Tb of stored data.
  • The performance benchmark will aim at delivering, on a 2-CPU server machine, using an Oracle 10 DB and a JBoss or BEA app server, with 3 concurrent users:
    • a 1.5 second response time when opening a space with 100 thumbnails
    • 0.5 seconds response time when doing an advanced search for a specific DICOM image, using 8 metadata attributes

Initial Project Scope


<outline the initial project scope>


Discussion of Design/Implementation Approach


  • Thumbnail View: based on the existing Forge Thumbnail project?
  • The code will be delivered back to Alfresco for integration into the Alfresco Core so that e.g.

Alfresco could deliver this as an add-on package in V2.2( using the packaging technology provided in Release 2.0.)


prerequisites


At the core of the dcm4che project is a robust implementation of the DICOM standard. The dcm4che-1.x DICOM toolkit is used in many production applications across the world, while the current (2.x) version of the toolkit has been re-architected for high performance and flexibility.


(http://www.luratech.com/download/index.jsp?OnlineShopId=0941DA3F3BA0BEA5064C14FF92BE4AEA&categoryKey=23)


for FireFox:




ImageMagick also has jpeg2000 conversion features


Document Links


the DynamicSpaces concept is really required for a convenient way to organize and present the content



Attachments

    Outcomes