White blood cell subtypes classification using capsule networks for automated hematological diagnostic

Key Points

• To develop a method using capsule networks for the classification of white blood cell subtypes aimed at improving diagnostic processes.
• Designed a capsule network architecture for white blood cell image analysis.
• Collected and preprocessed a dataset of hematological images for training and validation.
• Evaluated model performance against traditional classification methods.
• Achieved higher classification accuracy compared to conventional techniques.
• Demonstrated robustness in identifying a variety of white blood cell subtypes.
• Reduced diagnostic time significantly, indicating potential for clinical application.