Date on Master's Thesis/Doctoral Dissertation
Computer Engineering and Computer Science
Desoky, Ahmed H.
Committee Co-Chair (if applicable)
Microcirculation--Research; Image processing--Digital techniques; Computer vision
This thesis project consists of the development of a computerized image processing system to preprocess microvascular images of the mouse Latisimus Dorsi Muscle (LDM). This research has been conducted in association with the Division of Plastic and Reconstructive Surgery at the University of Louisville. The input to the system is a set of 35 or more overlapping microscopic fields or sub_images containing segments of the LDM microcirculation, the output is an adjacent single montage encompassing the entire LDM microvasculature. The developed system presents practical solutions to the problems of image registration, overlap resolution, and image segmentation, in addition, the results of this project include simple geometrical measurements of the microvascular system such as length, tortuosity, and directionality. In future projects, the measurements should be further analyzed to study the LDM microcirculation. The entire system was written in C/C++ languages, and the Bluestone's UIM/X GUI builder. The system was implemented on the Speed School's HP-9000 Unix system. The system also takes advantage of the image processing libraries (ImageMagick) available on the system.
Mansour, Mohamed S., "Preprocessing of microcirculatory images." (1995). Electronic Theses and Dissertations. Paper 1661.