Acute Lymphoblastic Leukemia Classification in Nucleus Microscopic Images using Convolutional Neural Networks and Transfer Learning

Abstract

Leukemia is a disease caused by the abnormal production of abnormal blood cells. In Acute Lymphoblastic Leukemia (ALL), lymphoblast cells do not develop into lymphocytes. To diagnose the disease, we need to differentiate between lymphocytes and lymphoblasts. However, lymphocytes and lymphoblasts have similar morphologies. Several studies using computer vision have been developed to distinguish lymphocytes from lymphoblasts. This study aims to compare deep and wide deep learning architectures to classify segmented blood cell images. The “deep” architecture employed in this study was DenseNet201, whereas the “wide” architecture was Wide-ResNet-50-2. We also employed ResNet50 as a baseline. This study utilizes transfer learning to reduce the training steps needed. In addition, we measured the impact of dataset preprocessing using histogram equalization on the classifier performance. We found that DenseNet201 model has the best performance with an AUC score of 86.73% and that histogram equalization makes the performance worse.