Hybrid classifier based life cycle stages analysis for malaria-infected erythrocyte using thin blood smear images

Malaria, being a life-threatening disease caused by parasites, demands its rapid and accurate diagnosis. In this paper, we develop a computer-assisted malaria-infected life-cycle stages classification based on a hybrid classifier using thin blood smear images. The major issues are: feature extraction, feature selection and classification of erythrocytes infected with different life-cycle stages of malaria. Feature set (134 dimensional features) has been defined by the combination of the proposed features along with the existing features. Features such as prediction error, co-occurrence of linear binary pattern, chrominance channel histogram, R–G color channel difference histogram and Gabor features are the newly proposed features in our system. In the feature selection, a two-stage algorithm utilizing the filter method to rank the feature, along with the incremental feature selection technique, has been analyzed. Moreover, the performance of all the individual classifiers (Naive Bayes, support vector machine, k-nearest neighbors and artificial neural network) is evaluated. Finally, the three individual classifiers are combined to develop a hybrid classifier using different classifier combining techniques. From the experimental results, it may be concluded that hybrid classifier formed by the combination of SVM, k-NN and ANN with majority voting technique provides satisfactory results compared to other individual classifiers as well as other hybrid model. An accuracy of 96.54 ± 0.73% has been achieved on the collected clinical database. The results show an improvement in accuracy (11.62, 6.7, 3.39 and 2.39%) as compared to the state-of-the-art individual classifiers, i.e., Naive Bayes, SVM, k-NN and ANN, respectively.