International Journal of Science, Engineering and TechnologyAuthors: Rajesh Vasant Jondhale, Nayan Sunil Sonawane, Yuvraj Rajendra Rasal, Nilesh Prakash Pawar, Prasad Yadav, Dr. Madhav J. Salunkhe
Abstract: Determining an individual's blood group is a vital step in medical diagnostics, traditionally conducted through laboratory-based blood tests. These methods, however, require blood samples, specialized equipment, and trained personnel, making them time-consuming and resource-intensive. This study introduces an innovative, non-invasive approach for blood group prediction using deep learning techniques applied to fingerprint images. Convolutional Neural Networks (CNNs) are utilized to extract and analyze distinct fingerprint features that correlate with blood group classification. The dataset comprises 6000 fingerprint samples, which are preprocessed using OpenCV techniques to enhance image quality and standardization. The CNN model is trained using the Adam optimizer over 25 epochs, ensuring effective learning while maintaining minimal training loss. Additionally, a web-based system is developed with Flask for front-end interaction and SQL Server for secure data management. This proposed framework offers a rapid, cost-efficient, and accessible alternative to conventional blood group detection methods, potentially benefiting remote healthcare services and emergency medical scenarios. The findings highlight the potential of biometric-based artificial intelligence in medical applications, paving the way for further research in non-invasive diagnostic techniques.
DOI: http://doi.org/
