Improved Classification of Malaria Parasite Stages with Support Vector Machine Using Combined Color and Texture Features

Malarial is a mosquito born deadly disease that quickly grows from person to person because of the infectious mosquito bite. Knowing accurately the developing stages of a parasite is critical for accurate drag selection for early recovery. However, limited study were found that dealt with the automated classification of malaria parasite stages. In this study, a supervised method for classifying malaria parasite stages from microscopy images has been proposed. To achieve the target, this method combines color and texture features with the support vector machine (SVM) classifier. Three texture features, namely histogram of oriented pattern (HOG), local binary pattern (LBP), Grey-level Co-occurrence Matrix (GLCM), and four color features, namely local color moments (StatMom) and color histograms (HSV, LAB, and YCrCb), have been considered. An experimental analysis with an unbalanced dataset of 46,978 single-cell thin blood smear images showed promising performances of the color features compared to the texture features. Using SVM classifier, the proposed color-texture feature (YCrCb_HOG) showed the highest classification accuracy (96.9%) on average, which exceeds the performance of a recently published method using HOG_LBP feature with the SVM (87.1%).