IDENTIFICATION SYSTEM FOR MALARIA PARASITE: PLASMODIUM VIVAX, USING MORPHOLOGICAL TECHNIQUES

Malaria is a parasitic disease caused by the bite of infected Anopheles mosquitoes.
The infection is caused by Plasmodium parasites transferred through the mosquitoes
bite. Currently the diagnosis system is performed manually by experts through
observation of blood smears under microscope. This diagnostic system is prone to
human error and very subjective. This project presents the development of a software
module to automatically identify Plasmodium Vivax (PV) in trophozoite stage using
morphological techniques. 350 PV images are taken using a video camera with 1000x
magnification that is attached to a microscope. An automated system to detect PV in
thin blood smear images is created using Matlab 2013a image processing toolbox.
The thin blood smear images will undergo pre-processing, segmentation and
classification stages. This system will focus on morphological techniques of dilation,
erosion and hole filling during the segmentation and the morphological hit-miss of
area for the classification stage. The process will classify the parasite in trophozoite
stage. 100 images are used for calculating the parasite area while 250 images are used
for final testing in detecting the parasite. The system produces 98.8% sensitivity,
99.90% accuracy and 99.96% specificity. The diagnosis time for one image is 0.2
seconds and 2 seconds for 25 images in one run. The result had shown that the system
is able to correctly classify the trophozoite parasites in the PV images in very short
time. With a reliable system, the time taken to diagnose malaria parasite can be
reduced and thus improves the current diagnosis process.