Effective Grayscale Conversion Method for Malaria Parasite Detection
Jun Yeon, Jong-Dae Kim, Yu-Seop Kim, Chan-Young Park, Hye-Jeong Song · Advanced science and technology letters · 2014
The main factor of malarial diagnosis is a Giemsa-stained blood smear image is effectively distinguish the stained pixels. The optimal grayscale conversion method for plasmodium-infected blood smear images is investigated in this paper, by comparing the classification performances of various grayscale conversion methods. We compare the performances of these methods by obtaining the area under the receiver operating curve (AUC) using 20 malaria- infected red blood cells (RBCs) and 20 normal RBCs. With the AUC value of 0.9225, the principal component analysis (PCA)-based grayscale conversion method outperformed all other methods.