Enhanced Blood Cell Classification Using Convolutional Neural Networks for Accurate Diagnosis of Haematological Conditions

Among haematological illnesses whose diagnosis and treatment depend critically on blood cell classification. Early discovery, good treatment planning, and monitoring of disease development all depend on accurate identification of blood cell types—especially malignant cells. We provide a CNN (Convolutional Neural Network) model in this work to categorize blood cells into four categories: Benign, Malignant_Pre-B, Malignant_Pro-B, and Malignant_Early_Pre-B. With a 3,242-image collection taken from Kaggle, our model shows quite remarkable overall accuracy of 98%. Particularly for the Malignant_Pre-B class, the proposed CNN architecture—which comprises of well-adjusted hyper parameters and network layers—effectively captures specific features across the several cell types, so producing good precision and recall metrics. The robustness and generalizing capability of the model are confirmed by a thorough investigation covering training and validation accuracy, loss curves, and a confusion matrix. This work emphasizes how CNNs can automatically recognize blood cells, hence enhancing clinical diagnosis speed and accuracy.