Blood Cell Image Classification based on CNN Algorithm

Abstract

The conclusion of blood-related infections includes the ID and portrayal of a patient's blood test. In that capacity, robotized techniques for recognizing and arranging the kinds of platelets have significant restorative applications in this field. Deep learning may solve this problem effectively. In the proposed system, convolutional neural network (CNN) is used for learning and detection. Most of existing research proposed were identifies blood cell class, whereas, this work aimed at blood cell type classification and disease identification as a combined model. This is achieved by training blood cell types as four classes separately and disease detection training (binary class of normal/cancer) using CNN algorithm. Experimental results show that CNN achieves more accuracy on training and validation set.