Publication

Abstract : Leukemia is a type of cancer that affects the body’s blood forming tissues, including bone marrow. It is a dangerous illness prevalent in children under the age of five. The present diagnosis includes microscopic examination of blood cells by the hematologist. Recently, deep learning methods are extensively employed in many medical imaging applications for the diagnosis of diseases. However, one of the key issues is the limited availability of microscopic images for training the models. To overcome this difficulty, transfer learning techniques are put forward. This paper presents a comparative analysis of different transfer learning models like Xception, Inceptionv3, DenseNet201, ResNet50, and MobileNet to detect acute lymphocytic leukemia (ALL) from blood smear cells. All models were trained on ALL-IDB2 dataset and achieved an accuracy of 87.97%, 88.92%, 88.92%, 95.28%, and 97.88%, respectively.