Customized CNN for Classification of Cho Cells

Abstract

Deep learning techniques such as Convolutional Neural Network have shown great promise in microscopic medical image classification tasks. Advancements in technology and research have indeed supported radiologists in detecting variabilities and other characteristics of microscopic medical images with improved accuracy and performance. In this paper we proposed variant of CNN for the Classification of CHO cells, evaluated on IICBU-2008 benchmark dataset and results are obtained for classification of CHO cells with 5 different classes with Accuracy 95.8 Precision 94.18, Recall 94.23 and F1 Score 94.2.