Classification of Blood Cells Using Deep Learning Models

Rabia Asghar, Sanjay Kumar, Abeera Mahfooz

Published 2023-08-11Version 1

Human blood mainly comprises plasma, red blood cells, white blood cells, and platelets. The blood cells provide the body's cells oxygen to nourish them, shield them from infections, boost immunity, and aid in clotting. Human health is reflected in blood cells. The chances that a human being can be diagnosed with a disease are significantly influenced by their blood cell type and count. Therefore, blood cell classification is crucial because it helps identify diseases, including cancer, damaged bone marrow, benign tumors, and their growth. This classification allows hematologists to distinguish between different blood cell fragments so that the cause of diseases can be identified. Convolution neural networks are a deep learning technique that classifies images of human blood cells (RBCs, WBCs, and platelets) into their subtypes. For this study, transfer learning is used to apply different CNN pre-trained models, including VGG16, VGG19, ResNet-50, ResNet-101, ResNet-152, InceptionV3 MobileNetV2 and DenseNet-201 to the PBC dataset's normal DIB. The overall accuracy achieved with these models lies between 91.375-94.72%. A novel CNN-based framework has been presented to improve accuracy, and we were able to attain an accuracy of 99.91% on the PBC dataset.