Automated Cells Counting for Leukaemia and Malaria Detection Based on RGB and HSV Colour Spaces Analysis

摘要

There are various types of diseases which are originated from the blood, for example leukaemia, malaria and anaemia. Leukaemia is a cancer which starts in blood forming tissues usually the bone marrow. On the other hand, malaria is transmitted through the bite of infected mosquito that carrying the Plasmodium parasite. Haematologists needs to perform the WBCs count in order to determine if a person has leukaemia and parasite count to check for the malaria density. However, the conventional procedure is very vulnerable due to human error and large time consumption. As a solution, this study proposes automated cells counting for leukaemia and malaria detection by analyzing the best colour component of RGB and HSV colour spaces. To obtain the cells counting result, there are several image processing steps to be implemented; (1) image acquisition by capturing the leukaemia blood samples using a computerized Leica DLMA 1200 digital microscope, (2) colour conversion from RGB to single colour component of RGB and HSV, (3) image segmentation using Otsu thresholding, (4) removing of unwanted regions and, (5) cells counting process. Overall, segmentation using green component of RGB colour space has proven to be the best in segmenting leukaemia images with 83.84% while saturation component of HSV colour space hold the highest accuracy for malaria images with 89.87%. Conclusively, this research is expected to help improving the detection phase of malaria and leukaemia diseases by overcome problems that been identify in this research.