Automated Malaria Parasitemia Estimation from Thin Blood Smear Images

Abstract

This paper presents an application of digital image processing in the medical field. Manual examination of the blood smear image for quantification of malaria is time consuming. Computer based approach not only atomizes the estimation process, but also increases the speed of computation, thereby reducing the efforts and the time. In this paper, an algorithm is proposed to automate estimating parasitaemia from thin blood smear images by image processing technique. The parasite infected and healthy (non infected) red blood cells are extracted and identified. The thresholding method is used for segmenting the red blood cells as well as cell parasites, in addition to morphological processing. Finally, the estimated parasitaemia was analyzed with manually acquired results. The analysis suggests that the accuracy of the method is comparable, promising and almost equal to manual method, computing result in fraction of seconds. The approach proposed is fast, accurate, effortless, simple and low cost.