2012 International Conference on Advanced Computer Science Applications and Technologies (ACSAT) (2012)
Nov. 26, 2012 to Nov. 28, 2012
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/ACSAT.2012.73
In this paper, we present a novel, fast, hybrid and bi-level segmentation technique uniquely developed for segmentation of medical images. Medical images are generally characterized by multiple regions, and weak edges. When regions in medical images are viewed as made up of homogeneous group of intensities, it becomes more difficult to analyze because quite often different organs or anatomical structures may have similar gray level or intensity representation. The complexity of medical imagery is well catered for in this technique by starting-out with multiple thresholding, applying similarity segmentation method, and resolving boundary problem with template matching technique, and then a region of interest (ROI) segmentation that involves finding the edges of the object of interest (OOI) at final stage. This technique can also be adapted to segmentation of non-medical images. A job is run using MATLAB and simple Grid computing as suitable environment.
computational complexity, edge detection, image matching, image representation, image segmentation, medical image processing
S. A. Hamed, A. A. Aboaba, O. O. Khalifa and A. H. Abdalla, "Hybrid and Multilevel Segmentation Technique for Medical Images," 2012 International Conference on Advanced Computer Science Applications and Technologies (ACSAT), Kuala Lumpur, 2013, pp. 442-445.