ISBI challenge 2019: Convolution neural networks for B-ALL cell classification

Abstract

We Participated in International Symposium on Biomedical Imaging (ISBI) 2019 challenge: Classification of Normal versus Malignant Cells in B-ALL White Blood Cancer Microscopic Images. Acute Lymphoblastic Leukemia (ALL) is a cancer of the lymphoid line of blood cells characterized by the development of large numbers of immature lymphocytes. In this paper, we present a Convolutional Neural Networks (CNNs) based solution for the challenge. We designed our solution with pretrained MobileNetV2 architecture as base classifier. Also, we employ transfer learning as the amount of labeled data is limited and ensemble the trained base classifier variants to reduce the generalization error on prospective data. With our final solution we secured second place in the challenge with F1 score of 0.8947.