Deep learning for classifying of white blood cancer

Abstract

Cell classification by computer assistant technique has become a popular topic recently. It is crucial to do early disease diagnosis with less cost. In order to find efficient, high accuracy computer assistant methods, The International Symposium on Biomedical Imaging (ISBI) held the classification of normal vs malignant cells (C-NMC) challenge 2019 for classifying B-ALL white blood cancer microscopic images. Our team deploys several state-of-the-art CNN architectures and uses an ensemble method named stacking to boost the final prediction performance, which helped to get the eighth place in the preliminary phase and tenth place in the final phase. Weighted F1 scores evaluated on our model were 0.8674 and 0.8552, for the preliminary and final phase, respectively. It is a relatively reasonable result in such an expertise required task.