Classification of Multiple White Blood Cell Combined Datasets by Using Feature Fusion-Based Ensemble Machine-Learning Models
Tansel Uyar · 2025
Convolutional neural networks and vision transformers are the most effective methods used in image classification problems. One of the current problems is the separation of white blood cells. In the contemporary context, the necessity for both expertise and an intensive workload in the separation of white blood cells has become increasingly evident. This necessitates the development of automated approaches capable of performing effective classification with a high degree of accuracy. Consequently, a proposed ensemble model classifier method integrates the two most effective approaches in literature and accommodates numerous decision makers. The proposed method yielded an approach that demonstrated both high generalization ability and effective discrimination. The performance of the final classifier was assessed using independent test data from diverse sources. This evaluation yielded highly successful outcomes, with an average accuracy of 98.88% and an average F1-score of 97.19%.