Automated determination of Plasmodium region of interest on thin blood smear images

The region of interest (RoI) has the most useful information in image processing since the targeted objects are covered in this area. By determining the precise position of RoI, a computer-based identification will be able to work more efficiently, to give a better contribution in system and to eliminate objects that may intrude overall process. In malaria disease, the existence of Plasmodium can be observed from patient's microscopic thin blood smear images. Having utilised the computer aided detection based on image processing, malaria disease can be detected earlier and more objective in order to support the final decision of paramedics. This study proposes a novel method to automatically determine the region of interest (RoI) on thin blood smear images for facilitating the process of Plasmodium parasite detection. The approach includes Otsu thresholding method, morphological operation and binary large object (BLOB) analysis. Evaluation results on 24 thin blood smear images show that the proposed method achieves the sensitivity and PPV rates of 87.5% and 75.7%, respectively. These successful results in automatically determine the RoI which contains the Plasmodium parasite indicate that the proposed method has a potential to be implemented in the development of a computer aided malaria detection system.