Convolutional Neural Network Approach for White Blood Cells Classification

Blood group detection plays a critical role in medical diagnostics, particularly in transfusion medicine and emergency healthcare. Traditional methods require invasive blood sampling, manual processing, and time-consuming laboratory tests. This paper proposes a novel deep learning-based system for blood group detection using both blood images and fingerprint images, offering a non-invasive alternative. The system utilizes the MobileNetV2 convolutional neural network (CNN) architecture, which is both lightweight and highly accurate. The first mode of detection uses blood images, achieving 100% accuracy in both training and validation. The second mode utilizes fingerprint images, achieving 94% training and 90% validation accuracy. By combining these two modes, the system offers flexibility, efficiency, and accessibility. It is designed using Python for backend development, Flask for the web framework, and HTML, CSS, and JavaScript for the frontend, ensuring a user-friendly interface. The proposed system not only improves the speed and accuracy of blood group detection but also makes it accessible to healthcare facilities with limited resources, reducing manual errors and labor-intensive processes.