Hybrid Whale Optimization Algorithm and CNN for White Blood Cell Image Classification

Abstract

This paper presents a novel blood type identification and classification system that combines the Whale Optimization Algorithm, Convolutional Neural Networks, and image processing. The suggested method aims to improve processing efficiency and accuracy, particularly in emergency medical situations like blood transfusions and surgery. The technique guarantees accurate and speedy blood type categorization by utilizing CNN's pattern recognition capabilities and optimizing feature selection through WOA. This method enhances patient safety and speeds up medical decision-making by reducing the chances of an immunological reaction brought on by blood type incompatibility. In order to identify distinctive features like texture, color, and shape, the study uses a BCCD dataset of blood cell images, preprocesses them to improve quality, and then applies CNN for pattern recognition. WOA improves the selection much more of these characteristics by imitating humpback whale hunting techniques, which results in increased precision and shorter processing times. According to experimental data, the suggested WOA-CNN system outperforms more conventional techniques like Decision Trees, Random Forests, Support Vector Machines, and Proposed Method (WOA-CNN) in terms of classification accuracy (98.6%). By reducing the immunological dangers connected to blood type incompatibility, this study demonstrates the possibility of combining deep learning and optimization algorithms to improve patient safety and medical diagnostics.