Deep Learning Model for the Automatic Classification of White Blood Cells.

Area of Science:

• Medical Imaging Analysis
• Computational Biology
• Artificial Intelligence in Healthcare

Background:

• Accurate blood cell counts are crucial for disease diagnosis.
• Traditional methods for blood cell analysis are invasive and equipment-intensive.
• Deep learning offers a promising alternative for automated analysis of blood cell images.

Purpose of the Study:

• To implement and evaluate a deep learning model for classifying white blood cells (WBCs).
• To optimize the DenseNet121 model using normalization and data augmentation for improved performance.
• To assess the model's effectiveness using various batch sizes and the Adam optimizer.

Main Methods:

• Utilized the DenseNet121 deep learning architecture.
• Applied preprocessing techniques including normalization and data augmentation.
• Trained and simulated the model on a dataset of 12,444 white blood cell images from Kaggle, testing with four batch sizes and the Adam optimizer over 10 epochs.

Main Results:

• Achieved a high classification accuracy of 98.84%.
• Demonstrated excellent performance metrics: 99.33% precision, 98.85% sensitivity, and 99.61% specificity.
• Identified batch size 8 as optimal for the DenseNet121 model's performance in this study.

Conclusions:

• The optimized DenseNet121 model provides a highly accurate and efficient method for WBC classification.
• This deep learning approach reduces the need for sophisticated equipment and time-consuming manual analysis.
• The developed model shows potential for clinical applications in automated WBC detection and disease diagnosis.