Digital medical images
acquired by medical imagers represent tremendous amount of data. The annual production of a single radiology department is tens of TB/year. Data semantics is important and data are often manipulated as correlated data sets. Medical data storage and retrieval therefore requires the manipulation of large volumes of data and meaningful associated metadata. Medical data manipulation is even more complex given the privacy constraint associated to individual data.
Given the amount of data produced in hospitals and the difficulty to interpret medical images, algorithms for medical image analysis, processing, and diagnostic assistance have been developed these last 15 years or so. Some of these algorithms have reached a high level of usability and proved to have a real impact in the clinical domain. However, their widespread adoption by clinicians is not realized yet.
|
|
Data and computing grids
are an opportunity to enlarge the impact of these image processing tools and to transfer this experimental research to clinical practice.
- Algorithm research and deployment
-
Availability of algorithms and datasets will ease the development, prototyping, and the validation of algorithms. Advanced users will be able to experiment and compare existing techniques on common data sets. Finally, grid-enabled algorithms will be accessible for clinical use.
- Image guided diagnosis and surgical planning
- Analysing large images at a sufficient speed to support interactive use requires substantial computing power. Combining the medical user expertise and the resource of the Grid in compute and data intensive tasks is a promising way to transfer experimental research first to clinical practice, and then to routine clinical practice.
- Augmented reality
- Pre-operative data, computed geometries, intra-operative images or the patient body itself can be combined to optimise a surgical intervention or therapy planning. Registration, intra-operative MRI for brain surgery, and multimode fusion are typical examples.
|
|
AGIR goals
are to define and validate
- New grid services
- Address some of the requirements of complex medical image processing and data manipulation application;
- New medical image processing algorithms
- Take advantage of the underlying grid infrastructure for compute and data intensive needs.
|
|
|
 |
|
AGIR expected results
- An integration testbed
- Improved knowledge of the requirements of medical applications, algorithms and IS towards grid middleware
- More powerful medical analysis algorithms
- Impact on national and international grid projects
- Clinical deployment experiments
|
|
|
|
