Publication: Pemprosesan imej berwarna dan pengiraan tubercle bacili
Loading...
Date
2006-05-01
Authors
Kassim, Mohd Hishamuddin
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
In Malaysia, the traditional method for the identification of Mycobacterium tuberculosis bacteria requires a trained technologist to manually examined sputum
specimen under 100 x 10 microscopes. This method is troublesome because it causes strain and eye fatigue and causes limits to only 25 specimens per day. It also takes 15 to 20 minutes to examine each specimen and it will cause error when reading results. As a solution, the development of a software system is proposed to automatically identify Mycobacterium tuberculosis cluster from digitized microscopic images of sputum specimens using C++ Builder Version 5 software. This software will count total bacteria in specimen and the result will determine whether the individual has tuberculosis disease or not. The technique used for counting bacteria is based on colour image processing. The red, green and blue (RGB) colour model will be used to process the colour image. The first step is to build the histogram for every colour component, and then the bacteria threshold value is identified to produce the threshold image. Next, filter image process will be done to eliminate the noise. Lastly, the bacteria will be count.