Image Classification of White Blood Cells Using Convolutional Neural Network
Juliet Cagampang, Ligaya Leah Figueroa · Atlantis Highlights in Computer Sciences/Atlantis highlights in computer sciences · 2025
Classification of white blood cells (WBCs) is a crucial process in medical diagnosis and research.Automated image classification of white blood cells using machine learning techniques provides faster and more accurate results compared to manual procedures.Convolutional Neural Networks (CNNs) are deep neural systems widely used in the medical classification tasks since they are excellent in feature extraction.This paper used the small Inception or MiniGoogleNet, a simple CNN, to classify five types of white blood cells, namely, basophils, eosinophils, lymphocytes, monocytes, and neutrophils.The quality of the dataset used to train, validate, and test the CNN classifier highly affects its performance.Hence, this paper presents a dataset preprocessing system that involves image processing to improve the quality of images and data augmentation using Albumentations to solve the problem of the highly imbalanced dataset.The model was trained from scratch using the Pytorch library and achieved an accuracy of 97.65%, recall of 94.12%, precision of 94.17%, and F1 score of 94.12%.