IDENTIFIKASI PARASIT MALARIA PLASMODIUM FALCIPARUM PADA SEDIAAN DARAH DENGAN PENDEKATAN SUPPORT VECTOR MACHINE

Abstract

Background: Identification of malaria is microscopic requiring special expertise and experience considerable health analyst. Factor errors that occur can be either an inability to recognize the parasite morphology and eyestrain factors in looking at morphology, this may impact
the diagnosis of significant errors. Morphology of Plasmodium falciparum is divided into three major stages: trophozoites, Schizonts and Gametocyter. From the results of research in the field of health shows, Schizonts found in peripheral blood showed a state of severe infections so it is an indication for rapid treatment measures. This study aims to identify the three-stage form of Plasmodium falciparum malaria parasites in digital images of blood preparations
which contain the parasite indicated.

Method: Before conducting the identification process, the first step of the analysis procedure conducted in this study is to conduct the separation of the object by using the segmentation k-means clustering. The second step, to extract features of image data to be tested. Feature extraction is used as an insert in the system to be constructed in this study using color features. The final step is to identify three forms of identification test stage of the malaria parasite plasmodium falciparum using a Support Vector Machine (SVM) method multiclass ones against ones.

Result: The results of this study using color as a characteristic feature of the input and identification using SVM can provide a success rate of 93.33% image data correctly.
Keywords: malaria, Plasmodium falciparum Morphology, k-means clustering, SVM multiclass method ones against ones.