MedGIFT

• About
• Partners
• Contacts
• Press Office
• Activities
• ARC Middleware
• Applications
• Demo videos
• Download and Use
• Standardization
• Documents
• Links
• News archive
• Jobs
• Internal Meetings
• Internal Wiki
• Home

• Grid Computing Inside Hospitals Using Virtualization Technology
• Demo medGIFT
• Taverna workflow management in medical imaging

Grid Computing Inside Hospitals Using Virtualization Technology

MedGIFT is a Grid infrastructure for medical imaging applications at the University Hospital in Geneva (HUG).

• Today's hospitals produce enormous amounts of images.
• The radiology division at Geneva Hospital produced more than 80000 images/day in 2008.
• Automatic analysis can aid diagnosis, but image retrieval and analysis is a computing intensive task that may be limited by available computing resources.

Deploy the Grid technology inside the Hospitals:

• The Grid technology can provide the required computing resources with both dedicated and existing machines
• Over 6000 desktop PCs are available on the network of the HUG
• To explore the idle computing resources, virtualization techniques (VMware) are deployed. The whole infrastructure is built up based on the ARC middleware.
• Three medical imaging applications (general content-based image retrieval, lung image retrieval, and fracture image retrieval) were gridified

However, there are many challenges. Hospitals deal with confidential material and firewalls block outbound connections. In addition, all hospital machines are Windows desktops only.

The MedGIFT solution exploits existing local resources with little communication overhead and avoids issues with firewalls by running only within the Hospital. The LRMS is run inside virtual machines in desktop PCs with the same VM image and software packages installed and configured on centrally controlled Linux inside Windows.

The application is based on Gridified GNU Image Finding Tool (GIFT).

Demo medGIFT

Follow this link for an online medGIFT demo: medGIFT

The application helps to search through medical images, and allows users to refine the search by marking results as useful or not, before running the search again. Finally, users can retrieve information about the desired results and the cases they represent, for example how they were treated.

Taverna workflow management in medical imaging

Taverna is a user-friendly Graphical User Interface (GUI) which helps users to quickly adapt Grid computing without much knowledge of technical details. It allows developers to quickly adapt their applications to Grid execution using a GUI workflow management. An ARC plug-in takes care of job submission, job status checking, as well as result collection. The figure below shows an example of using the ARC Grid via this tool.