การออกแบบและพัฒนาโปรแกรมเพื่อวัดโครงสร้างของผนังหลอดเลือด / อุรีรัฐ วัฒนชนม์ = A design and development of a program for measuring vascular wall structure / Ureerat Wattanachon
Designs and developes a program for measuring vascular wall structure by using digital image processing and computer graphics to help analyzing images. Various different blood vessel shapes, slide preparation and image acquisition processes can all cause blood vessel images to vary widely. The designed measuring program, hence, has some limitations that users must be aware of, and they must select a suitable working mode. The measuring program is designed to work in 3 modes: automatic, semi-automatic and manual. Appropriate measuring points on vascular walls are automatically selected in the automatic mode. Vascular wall structure, which includes the vascular wall thickness, is calculated. With semi-automatic mode, users can choose to delete, to add or to change the measuring points selected from automatic mode. Users can freely select measuring points to measure vascular wall thickness in the manual mode. In this research, the measuring program has been evaluated with 22 images of both circular rings and elliptical rings of 1 cm. and 2 cm. thick. The thicknesses resulting from the program are 0.994+-0.004 cm. (n = 313) and 1.981+-0.025 cm. (n = 242) for the 1 cm. and 2 cm., respectively. The program has been tested with 27 blood vessel images, when compared to human measuring results, shows that it provides more reliable measurements as many more measuring points are used in the program, especially for the distinct vascular walls images. A practical use at the Microcirculation Lab, Department of Physiology, Faculty of Medicine, Chulalongkorn University, shows that the program makes vascular wall measuring process work much more convenient. The results are more statistically efficient when compared to those done by human because human cannot select as many measuring points. The program also reduces human's uncertain measuring problems which cause measuring results to vary widely.
