Advanced Blood Cell Image Classification Using a Convolutional Neural Network Approach

Early blood cell type classification provides essential diagnostic power in hematological disorders by enabling doctors to perform targeted treatment at the earliest possible stage. The study presents a novel sequential Convolutional Neural Network (CNN) architecture that automatically classifies blood cell images into Eosinophil, Platelet, Erythroblast, Monocyte, Basophil, and Lymphocyte categories. Medical facilities need reliable automated systems to properly support their diagnostic workflows because precise early cell recognition plays a critical role in healthcare. The available dataset contains 10,868 labeled images distributed evenly into six classification groups. A robust and balanced analysis required dividing the dataset into training (80%), validation (10%), and testing (10%) partitions. The CNN model received training through 8,694 images before validation with 1,087 examples for overfitting prevention, and finally tested with 1,087 independent images. The proposed model displayed outstanding performance by reaching 98% accuracy in classifying all blood cells with a high level of generalized accuracy. The performance evaluation showed high stability through precision, recall, and F1-score metrics, which exceeded 0.95 values for each class type. The proposed CNN model demonstrates high effectiveness and reliability at distinguishing blood cell types, making it a strong candidate for clinical diagnostic tools and early hematological disease detection systems.