Diagnosis of Malaria Using Wavelet Coefficients and Dynamic Time Warping

Abstract

Malaria remains one of the world’s most deadly infectious disease and arguably, the greatest menace to modern society in terms of morbidity and mortality. Demand of time is to identify the methods, which can diagnose the disease and are economical and accurate. In this work, we have proposed a technique based on scientific computations. The aim is to provide a method which can do classification of Red Blood Cells (RBCs) into the malaria infected and healthy ones. The images of RBCs are captured using low cost digital holographic interferometric microscope (DHIM), these images are pre-processed for noise removal and then used for computing the discrete wavelet coefficients (DWC). The DWC are then used, as feature vectors, in dynamic time warping algorithm to find similarities or dissimilarities, between the RBCs. This allows the classification of RBCs into healthy or malaria infected. The proposed method shows excellent results for classification of RBCs for diagnosis of disease. Such automated detection methods, using low cost DHIM devices and minimal human intervene, will be good aid to medical area and are demand of time.