A Customized Efficient Deep Learning Model for the Diagnosis of Acute Leukemia Cells Based on Lymphocyte and Monocyte Images

The production of blood cells is affected by leukemia, a type bone marrow cancer or cancer. Deoxyribonucleic acid (DNA) related to immature cells, particularly white and damaged in various ways this disease. When radiologist involved diagnosing acute leukemia the diagnosis time consuming needs provide better accuracy. For purpose, many types research have been conducted for automatic leukemia. However, these studies low detection speed Machine learning artificial intelligence techniques are now playing an essential role medical sciences, detecting classifying leukemic cells. These methods assist doctors diseases earlier, reducing their workload possibility errors. This aims design deep model with customized architecture using images lymphocytes monocytes. study presents novel dataset containing Acute Lymphoblastic Leukemia (ALL) Myeloid (AML). new has created assistance experts help scientific community its efforts incorporate machine into research. Increasing scale achieved Generative Adversarial Network (GAN). proposed CNN based on Tversky loss function includes six convolution layers, four dense Softmax activation classification images. 99% accuracy rate types, including ALL AML. Compared previous research, network provides promising performance terms accuracy; results, can be used specialists practical applications.