Data grid for large-scale medical image archive and analysis

H. K. Huang, Aifeng Zhang, Brent Liu, Zheng Zhou, Jorge Documet, Nelson King, Wing Chi Chan

Research output: Chapter in book / Conference proceedingConference article published in proceeding or bookAcademic researchpeer-review

31 Citations (Scopus)

Abstract

Storage and retrieval technology for large-scale medical image systems has matured significantly during the past ten years but many implementations still lack cost-effective backup and recovery solutions. As an example, a PACS (Picture Archiving and Communication system) in a general medical center requires about 40 Terabytes of storage capacity for seven years. Despite many healthcare centers are relying on PACS for 24/7 clinical operation, current PACS lacks affordable fault-tolerance storage strategies for archive, backup, and disaster recovery. Existing solutions are difficult to administer, and often time consuming for effective recovery after a disaster. For this reason, PACS still encounters unexpected downtime for hours or days, which could cripple daily clinical service and research operations. Grid Computing represents the latest and most exciting technology to evolve from the familiar realm of parallel, peer-to-peer, and client-server models that can address the problem of fault-tolerant storage for backup and recovery of medical images. We have researched and developed a novel Data Grid testbed involving several federated PAC systems based on grid computing architecture. By integrating grid architecture to the PACS DICOM (Digital Imaging and Communication in Medicine) environment, in addition to use its own storage device, a PACS also uses a federated Data Grid composing of several PAC systems for off-site backup archive. In case its own storage fails, the PACS can retrieve its image data from the Data Grid timely and seamlessly. The design reflects the Globus Toolkit 3.0 five-layer architecture of the grid computing: Fabric, Resource, Connectivity, Collective, and Application Layers. The testbed consists of three federated PAC systems, the Fault-Tolerant PACS archive server at the Image Processing and Informatics Laboratory, the clinical PACS at Saint John's Health Center, and the clinical PACS at the Healthcare Consultation Center II, USC Health Science Campus. In the testbed, we also implement computational services in the Data Grid for image analysis and data mining. The federated PAC systems can use this resource by sharing image data and computational services available in the Data Grid for image analysis and data mining application. In the paper, we first review PACS and its clinical operation, followed by the description of the Data Grid architecture in the testbed. Different scenarios of using the DICOM store and query/retrieve functions of the laboratory model to demonstrate the fault-tolerance features of the Data Grid are illustrated. The status of current clinical implementation of the Data Grid is reported. An example of using the digital hand atlas for bone age assessment of children is presented to describe the concept of computational services in the Data Grid.
Original languageEnglish
Title of host publicationProceedings of the 13th ACM International Conference on Multimedia, MM 2005
Pages1005-1013
Number of pages9
DOIs
Publication statusPublished - 1 Dec 2005
Externally publishedYes
Event13th ACM International Conference on Multimedia, MM 2005 - Singapore, Singapore
Duration: 6 Nov 200511 Nov 2005

Conference

Conference13th ACM International Conference on Multimedia, MM 2005
Country/TerritorySingapore
CitySingapore
Period6/11/0511/11/05

Keywords

  • Bone age assessment of children
  • Computational services
  • Data grid
  • Fault-tolerance archive
  • Grid Computing
  • Image analysis
  • Image data mining
  • PACS

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design
  • Computer Vision and Pattern Recognition
  • Human-Computer Interaction
  • Software

Fingerprint

Dive into the research topics of 'Data grid for large-scale medical image archive and analysis'. Together they form a unique fingerprint.

Cite this