BioIDScan: Fingerprint-Based Blood Group and Pattern Analysis Using Machine Learning

This paper presents BioIDScan, a biometric classification system designed to predict fingerprint patterns and human blood groups from grayscale fingerprint images using deep learning. Convolutional Neural Networks (CNNs), trained on the NIST SD4 dataset for fingerprint patterns and a Kaggle dataset for blood group classification, are employed. Preprocessing is handled via OpenCV to ensure consistency. TensorFlow/Keras models are integrated into a React front end and Flask back end, enabling realtime prediction. Results show 94.3% accuracy in fingerprint pattern recognition and 82.7% in blood group prediction. This study demonstrates the potential of integrating biometric authentication with non-invasive health screening.