Tomek links-based SMOTE method for class imbalance in blood cell classification with dual path sliding window attention model

Key Points

• Improved classification accuracy is achievable with the Tomek links-based SMOTE method, addressing class imbalance issues.
• The SMOTE method successfully enhances performance metrics against a baseline accuracy in blood cell classification systems.
• Analysis utilized a dual path sliding window attention model to refine classification, enabling the focus on important input features.
• Findings indicate that addressing class imbalance through advanced techniques is crucial for enhancing analysis in medical settings.