Tidarat Katsanook, Chalermrat Nontapa, and Kornprom Pikulkaew
Published in Data Science and Engineering (DSE) Record 2025 Vol. 6 No. 1 pp. 413-432
Abstract
White blood cell (WBC) classification plays a pivotal role in diagnosing and monitoring various medical conditions, particularly hematological and immune-related disorders. This study explores the application of machine learning (ML) and deep learning (DL) techniques to classify WBCs, leveraging their potential to enhance diagnostic precision and efficiency. Using a dataset of 50,000 2D images from the University of North British Columbia, we develop and evaluate models for categorizing WBCs into four key types: eosinophils, lymphocytes, monocytes, and neutrophils. The proposed methodology integrates data augmentation, feature extraction, and advanced classification algorithms, including Convolutional Neural Networks (CNNs) and other statistical approaches. Performance metrics such as accuracy, precision, recall, and F1-score guide the optimization of model architecture and training processes. Experimental results demonstrate the effectiveness of the developed models in achieving high classification accuracy, offering a reliable and automated tool for WBC identification. This research underscores the potential of AI-driven solutions to improve clinical workflows, particularly in resource-limited settings, by providing accessible and cost-effective diagnostic support.