On classification of the cytological images of leukocytes with transfer deep learning models

Eugene Yu. Shchetinin, Leonid Sevastianov, Anastasia Glushkova, Anastasia Demidova
White blood cells, also known as white blood cells (WBC), play an important role in protecting the human body from foreign intruders, including bacteria and viruses, and dangerous diseases caused by them. The article implements a computer approach to the classification and detection of white blood cells on cytological images of blood cells using deep learning methods. A LeucoCyteNetv2 blood image classification model is proposed, the architecture of which includes SeparableConv2D convolutional layers separated by depth. The developed model classifies leukocytes with an accuracy of 98.86\%, which allowed us to assert the possibility of their use as an auxiliary tool for hematological blood analysis.