Classification of lymphocytes, monocytes, eosinophils, and neutrophils on white blood cells using hybrid Alexnet-GoogleNet-SVM

Abstract White blood cells (WBC), which form the basis of immune system, protect body from foreign invaders and infectious diseases. While number structural features WBCs can provide important information about health people, ratio subtypes these observable deformations are a good indicator in diagnostic process. The recognition type lymphocytes, neutrophils, eosinophils, basophils monocytes is critical. In this article, Deep Learning based Hybrid CNN (Convololutional Neural Network) model proposed for classification monocytes, neutrophils WBCs. presented on pretrained Alexnet Googlenet architectures. feature vector last pooling layer both architectures has been merged, resulting classified by Support Vector Machine. To determine superiority method, was also performed compared using Googlenet. Alexnet-Googlenet-SVM provides higher accuracy than method tested with WBC images Kaggle LISC database. Accuracy F1-score were 99.73%, 0.99 98.23%, 0.98 data sets, respectively.