A NOVEL DETECTION AND CATEGORISATION OF WHITE BLOOD CELLS USING A DUAL CONVOLUTIONAL NEURAL NETWORK APPROACH

Abstract

Leukocyte detection and segmentation (WBCs) are a crucial stage in haematology imaging; a significant part in the diagnosis and management of serious illnesses is played by cell categorisation, particularly that of Leukocytes. White blood cells are also referred to as leukocytes which act as a foundation of the immune system. There are different approaches for the detection of WBC, but there are fewer papers which classify the type of WBC. The different approaches include Deep learning techniques and Support Vector Machine (SVM). The proposed methodology used deep learning strategies using the Convolutional Neural Network (CNN). Utilising blood cell imaging and the White blood cells (WBC), sequential image cropping technique and created a dataset of WBC. Then using CNN, the classification of images and type of WBC is determined as output. Through our CNN model, we can overcome the data augmentation by contemplating the blood cell images dataset and generate an accuracy of 99.191% by producing relevant results.