Efficient Leukemia Detection From Blood Images Using MobileNet
Sivamurugan Chelladurai, Madala Chaitanya Prabhu, Tuttaganti Saisree, Jampana Manojna, K. Manoj Kumar, D Chaithra · 2025
White blood cells production is irregular in acute lymphoblastic leukemia (ALL), a potentially fatal illness that effects the bone marrow and blood. As a result of these harmful cells displacing healthy ones, the immune response is compromised and the body gets more at risk for infections. The Suggested application makes clear how important and precise identification is to successful treatment and patient outcomes. By identifying leukemia and its phases through the analysis of microscopic pictures of blood samples, the suggested model provides a reliable answer. Through the use of a Convolutional Neural Network (CNN) architecture, the model is capable to identify whether leukemia is early pre-B and pro-B. It shows as benign and gives further details about the sample if no tumors are found guaranteeing through findings (depending on the Accuracy). Additionally, if a user uploads an inaccurate or non-blood sample image, the system has an error-handling function that ensures prompting the user to provide the proper image. In order to help medical professionals better explain the disease’s stages to their patients and help them make more educated treatment decisions, the model also offers basic information about each stage. By removing the laborious and error-prone manual examination procedure, this application is specifically made for healthcare professionals to enable quick and accurate diagnosis and, ultimately, improves patient care outcomes.