Title: Detecting acute leukaemia in blood slides images using a CNN ensemble

Authors: Maíla L. Claro; Rodrigo M.S. Veras; Luis H.S. Vogado; André M. Santana; Vinicius P. Machado; Justino D. Santos; João Manuel R.S. Tavares

Addresses: Departamento de Computação, Universidade Federal do Piauí, Teresina, PI, Brazil ' Departamento de Computação, Universidade Federal do Piauí, Teresina, PI, Brazil ' Departamento de Computação, Universidade Federal do Piauí, Teresina, PI, Brazil ' Departamento de Computação, Universidade Federal do Piauí, Teresina, PI, Brazil ' Departamento de Computação, Universidade Federal do Piauí, Teresina, PI, Brazil ' Departamento de Informática, Instituto Federal do Piauí, São Raimundo Nonato, PI, Brazil ' Departamento de Engenharia Mecânica, Faculdade de Engenharia, Instituto de Ciência e Inovação em Engenharia Mecânica e Engenharia Industrial, Universidade do Porto, Porto, Portugal

Abstract: Leukaemia is a disease that has no defined etiology and affects the production of white blood cells in the bone marrow. Young cells or blasts are produced abnormally, replacing normal blood cells (white, red blood cells, and platelets). Consequently, the person suffers problems in transporting oxygen and infections combat. Acute leukaemia is a particular type of leukaemia that causes abnormal cell growth in a short period, requiring a quick start of treatment. Classifying the types of acute leukaemia in blood slide images is a vital process, and a system of assisting doctors in selecting treatment becomes necessary. This article presents an ensemble approach composed of three convolutional neural networks (CNNs) - AlertNet-RWD, ResNet50 and InceptionV3. These CNNs, individually, demonstrated effectiveness in differentiating blood slides with acute lymphoid leukaemia (ALL), acute myeloid leukaemia (AML), and healthy blood slides (HBS). We verified that the union of these three well-known CNNs improves the hit rates of current techniques from the literature. The experiments were carried out using 18 public data sets with 3,293 images, and the proposed CNN ensemble achieved an accuracy of 96.17%, and precision of 96.38%.

Keywords: acute leukaemia diagnosis; model ensemble; convolutional neural network; CNN.

DOI: 10.1504/IJICA.2023.129355

International Journal of Innovative Computing and Applications, 2023 Vol.14 No.1/2, pp.22 - 33

Received: 23 Sep 2020
Accepted: 16 Mar 2021
Published online: 07 Mar 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article