A medical image processing system based on a hybrid computer architecture

  • Keith O'Connor

    Student thesis: Master's Thesis


    In many clinical situations it is necessary to view or measure internal parts of a patient's anatomy. The less invasive the method used to facilitate the viewing or measurement the better. This thesis documents the research undertaken in order to facilitate the development of an image processing system which helps meet some of the needs of the clinical situation. In particular, the system described facilitates the display and interactive segmentation of medical images. Once images have been segmented into meaningful regions of interest, the constituent parts can be measured and quantified to help in clinical diagnosis.

    The developed system is based on parallel architecture which enables it to function at reasonable speed. The user interface has also been designed to be user friendly and, is for the most part, reasonably intuitive. In addition to the normal image processing facilities, the developed system provides the user with the means to store information relating to the images obtained. This information has obvious value to clinicians when they come to view images at a time later to when they were first obtained.

    The system has been applied successfully to Magnetic Resonance images but is suitablefor use with other imaging modalities.
    Date of AwardJan 1998
    Original languageEnglish
    SupervisorAndrew Ware (Supervisor)


    • image processing system
    • parallel processing architecture
    • image segmentation
    • 2D medical images
    • medical imaging
    • Magnetic Resonance images
    • Computer systems for medical imaging

    Cite this