Affiliations: [a] Department of Convergence Software, Hallym University, Korea | [b] Bio-IT Research Center, Hallym University, Korea
Correspondence:
[*]
Corresponding author: Hye-Jeong Song, Bio-IT Research Center, Hallym University, 1 Okchon-dong, Chunchon-city, Gangwon-do, 200-702, Korea. E-mail:[email protected]
Abstract: Malaria must be diagnosed quickly and accurately at the
initial infection stage and treated early to cure it properly. The malaria
diagnosis method using a microscope requires much labor and time of a
skilled expert and the diagnosis results vary greatly between individual
diagnosticians. Therefore, to be able to measure the malaria parasite
infection quickly and accurately, studies have been conducted for automated
classification techniques using various parameters. In this study, by
measuring classification technique performance according to changes of two
parameters, the parameter values were determined that best distinguish
normal from plasmodium-infected red blood cells. To reduce the stain
deviation of the acquired images, a principal component analysis (PCA)
grayscale conversion method was used, and as parameters, we used a malaria
infected area and a threshold value used in binarization. The parameter
values with the best classification performance were determined by selecting
the value (72) corresponding to the lowest error rate on the basis of cell
threshold value 128 for the malaria threshold value for detecting
plasmodium-infected red blood cells.
