White Blood Cell Classification Using Graph Attention Neural Network 

The classification of white blood cell images plays a vital role in hematologic diagnosis and disease monitoring. However, existing deep learning approaches still face challenges such as overlapping cells, inconsistent morphology, and variations in image quality due to lighting or staining inconsistencies, which limits robustness in clinical applications. To address these issues, this study proposes a novel hybrid deep learning framework, YOLO- GTNet, that integrates YOLOv11 for high-precision object detection, Graph Attention Networks for modeling spatial interactions, and a Transformer-based head for robust classification. The model is trained on the large-scale VSB - WBC dataset (16,027 images) and evaluated using accuracy and F1-score. The proposed method achieves an average accuracy of 98.46% and an F1-score of 98.44% across nine classes of WBC images, outperforming conventional models, including Convolutional Neural Networks, CheXNet, and Faster R-CNN, indicating robust performance across white blood cell image subtypes. To the best of our knowledge, this study is among the first to combine YOLOv11, Graph Attention Network, and a Transformer-based head for WBC classification, enhancing spatial reasoning and overall performance. The results demonstrate the potential of the proposed framework for integration into real-time clinical decision-support systems. Further clinical validation is encouraged to confirm its effectiveness in real-world diagnostics. Supplementary Material File (white blood cell classification using graph attention neural network.pdf) - Download - 1.59 MB Information & Authors Information Version history Copyright This work is licensed under a Creative Commons Attribution 4.0 International License

Authors Metrics & Citations Metrics Article Usage 254views 153downloads Citations Download citation Minh Ly Duc, Matej Sindelar, Kiet Vo Thanh, et al. White Blood Cell Classification Using Graph Attention Neural Network . Authorea. 05 December 2025. DOI: https://doi.org/10.22541/au.176496687.72600015/v1 DOI: https://doi.org/10.22541/au.176496687.72600015/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu.