Acute Lymphoblastic Leukemia (ALL) Classification from Microscopic Images Using CNN Model ‘ALLNet’

Leukemia is a common blood cancer that affects people of all ages and is one of the major causes of death, presenting a serious threat to world health. In this study, we have proposed a Convolutional Neural Network (CNN) model called’ ALLNet ‘ for the precise diagnosis and classification of acute lymphoblastic leukemia (ALL) in human bone marrow images. The model distinguishes between immature lymphoblasts and normal cells. Additionally, we performed a comparative analysis with a few unique CNN models and six pre-trained models, utilizing both Transfer Learning and Fine-Tuning techniques. The ‘ALLNet’ model achieved a high validation accuracy of 93.05%, a test accuracy of 93.43%, and a corresponding F1-score of 93.43%, with Roc curve area of 0.97. In terms of accuracy, precision, recall, and F1-score, our proposed model performed better than all the models included in this study as well as other models that already exist. The ‘ALLNet’ demonstrates effectiveness and promise as a reliable method for classifying leukemia. The ‘ALLNet’ is a robust tool for medical professionals to accurately diagnose and classify acute lymphoblastic leukemia (ALL) using bone marrow images. It offers the potential to improve patient care by aiding in early and accurate diagnosis, leading to timely treatment interventions. Additionally, it showcases the capability of CNN models in medical image analysis and classification tasks, contributing to advancements in healthcare technology.