Abstract
Leukemia is a malignant cancer that occurs when the body begins to have a sudden increase in the number of white blood cells. Malignant blood cancer has many different types. Treatment of leukemia is also extremely complicated, expensive and has a very high mortality rate. Blood cell classification during testing is very important to diagnose and detect diseases promptly. It also become a challenge for doctors to diagnose blood cancer. In this thesis, we propose a method to automatically classifying blood cells using a deep learning network, CNN-LSTM. The proposed method classifies blood cells including red blood cells and white blood cells on blood cell images after staining. Experimental results demonstrate that the proposed method using CNN-LSTM network model achieves an accuracy of 85.41%, higher than that of CNN at 79.635%. This method effectively supports doctors in identifying blood cell types for rapid detection and timely treatment of patients with leukemia.