A Comparison Study Between Deep Learning and Conventional Machine Learning on White Blood Cells Classification

Abstract

White blood cells have a major role for human health. Classification of white blood cells can help psychiatrist to diagnose disease caused by abnormality in white blood cells. Diagnosis by human is subjective and prone to error. In this study, Deep Learning method is proposed to classify the two most common type of white blood cells. The result from Deep Learning also compared with three conventional machine learning methods. The conventional machine learning refer to method that cannot learn directly from raw data. Those conventional machine learning were Multi Layer Perceptron, k-Nearest Neighbour and Support Vector Machine. There were 9 texture features utilized by conventional machine learning. Deep Learning outperformed all of those three conventional machine learning. The best achieved accuracy by Deep Learning was 0.995.