Improved Convolutional Neural Network for Classification of White Blood Cells
The Classification of White Blood Cells (WBC) is a crucial task as the toll of WBC gives valuable information about the human health as their primary task is to build the immune system by fighting the foreign objects in the human body such as viruses and a certain type of bacteria, thus they prevent the body by falling into ill. The fundamental ambition of this paper is to create a classification method for accurate and efficient segmentation of white blood cells by applying a deep Convolutional Neural Network (CNN) model approach. We employed CNN architecture in this paper because of its accuracy and its automatic detection of important features without human intervention. Two real medical hyperspectral image sets show experimental results that cell classification using CNNs is efficient. In comparison, the proposed approach, employing spatial and spectral features jointly, will achieve better classification efficiency compared to standard support vector machines (SVM) by demonstrating the enormous potential of the CNN-based approach for accurate medical hyper-spectral data classification. We employed this architecture on Kaggle dataset of “Blood Cell Images”.