Classification of White Blood Cells using Convolutional Neural Network with Data Augmentation

Luong^1*

¹ Department of Electronic Technology and Biomedical Engineering, Hanoi University of Science and Technology, Vietnam

² Hue Central Hospital, Viet Nam

^*Corresponding Author

Duong Trong Luong

Department of Electronic Technology and Biomedical Engineering

Hanoi University of Science and Technology

Vietnam

Email: luong.duongtrong@hust.edu.vn

Abstract

Nowadays, along with the development of science and technology, the classification of white blood cells (WBC) can assist in the diagnosis of leukemia. Manual blood cell classification is often not very accurate, time-consuming, laborious, and costly. Moreover, the risk of manual blood collection can be the transmission of infectious diseases such as HIV/AIDS, which causes unnecessary harm. Therefore, developing an automated system for blood cell classification will improve safety and save time and money. In this work, we proposed the classification method of white blood cells using Convolutional Neural Network (CNN) with Data Augmentation. To test the effectiveness of the proposed white blood cell classification system, a total of 10,299 white blood cells images from CellaVision was used. The experimented results of the proposed method have achieved 96% accuracy and compared to other published methods.

Keywords: White blood cells, Classification, Convolutional Neural Network, Data Augmentation, Blood disorder

Download the full article (PDF)