Automated detection of Southeast Asian Ovalocytosis (SAO) obtained from thin blood smear microphotographs

Southeast Asian Ovalocytosis (SAO) is an erythrocyte disorder, which is characterized by oval-shaped cells with one or two transverse ridges or a longitudinal slit on blood smears. To check the SAO using thin blood smear requires expert manpower and is time consuming. This research was aimed to automatically detect the SAO from thin blood smear images. Digital images were acquired using a digital camera connected to a light microscope. Images underwent gray-scale conversion to decrease representation of images. Otsu?s Method was implemented to separate blood components and the background. All clumps present in the image were then processed to be separated using recursive bottleneck detection algorithm. The major and minor axes of each extracted erythrocyte were then obtained for the calculation of ratio. The ratio determined whether the erythrocyte is normal or oval, and a count was kept between them. The percentage of oval erythrocyte over total erythrocyte was used as the determination of SAO positive or negative. The major and minor axes of the erythrocytes were successfully detected with error rate lower than 1%. Thin blood smear sample, prepared and provided by Eijkman Institute for Molecular Biology Indonesia, were used to build and test of the proposed algorithm.