Malariaparasitemia is used for measuring the degree of infection by detecting the plasmodium (commonly known as the malaria parasite) in the blood of an infected patient. The most commonly used method of malarial diagnosis is counting the number of malariainfected red blood cells in a Giemsa-stained blood smear by using a microscope. This method requires expert knowledge and is prone to inter-investigator variability. Therefore, a number of studies have been conducted on automated classification techniques that can measure malarial infection rapidly and accurately. In order to detect malaria parasites by analyzing plasmodium-infected blood smear images, conversion processing is required to make the images insensitive to the luminance contrast of microscopy and staining intensity. This paper aims to identify a grayscale conversion method optimal for plasmodium-infected blood smear images by comparing the performances of various grayscale conversion methods. The grayscale conversion methods selected for the comparison are colorimetric conversion, luma coding, conversion using the green channel only, and principal component analysis (PCA)based conversion. We used 20 malaria-infected red blood cells and 20 normal red blood cells to compare the performances of these methods by obtaining thearea under the receiver operating curve (AUC) as the minimum histogram intra-class variance value for each cell image. With the AUC value of 0.9225, the PCA-based grayscale conversion method outperformed all other methods.