Classification, Segmentation and Quantification of acute lymphoblastic leukemia cells in the bone marrow
Theou Jeffine, S. Usha Kiruthika, J. Joshan Athanesious · 2024
Cancer is a type of disease in which abnormal cells flourish and penetrate adjacent tissues, then spread to other parts of the body via the blood and lymph system. This abnormality results in DNA alterations in the bone marrow, which give rise to malignant cells in haematologic malignancies like leukaemia. This results in the abnormal production of white blood cells, where cancerous cells proliferate more quickly than healthy cells, compromising the ability of healthy cells to carry out their regular tasks, including fighting infections, distributing oxygen, and clotting blood. The primary objective was to create an efficient model that could classify the given microscopic image of bone marrow containing cancerous cells or not. If the image is classified as cancerous then the model will segment the cancer cells over the normal cells, in addition the model will quantify the presence of malignant cells to measure the spread in the blood. Segmentation and quantification along with the classification will provide better explainability to the model, thus to achieve this both the image processing techniques and the deep learning model are used to develop the Classification, Segmentation and Quantification (CSQ) of acute lymphoblastic leukemia cells in the bone marrow. The model is evaluated using the metrics like accuracy, precision, recall, specificity and F1score. Experimental results showed that the model performs well when compared to the existing methods.