MUMBAI, India, June 26 -- Intellectual Property India has published a patent application (202641053279 A) filed by Varun Kumar B; Salmaan M; Sanjay R; Sreenath G; and Subash S B on April 27, 2026, for Blood Group Prediction Using Fingerprint Analysis.
Inventors include Varun Kumar B; Salmaan M; Sanjay R; Sreenath G; and Subash S B.
The application for the patent was published on June 19, 2026, under issue no. 25/2026.
Abstract: This project presents the design and development of Non-Invasive Blood Group Prediction Using Fingerprint Analysis, an Al-powered web application built with Python/Flask. The application predicts ABO/Rh blood groups (A+, A-, B+, B-, AB+, AB-, O+, O-) from fingerprint images through intelligent image processing and CNNbased classification using EfficientNetBO, achieving 90.33% test accuracy on the SOCOFing dataset. The system focuses on extracting unique ridge features such as loops, whorls, and arches, and mapping them to blood group classifications using trained deep learning models. It leverages modem technologies to deliver a fast, seamless demo experience for academic research: Python 3.10+, TensorFlow/Keras for deep learning, OpenCV for preprocessing, Flask for backend API and template rendering, HTML/CSS/JavaScript for a cyberpunk-themed interactive frontend, and Matplotlib/Seabom for visualization of model performance. The application follows a structured pipeline including image preprocessing, feature extraction, model inference, and result visualization, ensuring both accuracy and efficiency. The modular architecture allows easy scalability, future upgrades, and integration with advanced frameworks or cloudbased deployment systems. Users can upload fingerprint scans to receive instant blood group predictions, along with ridge pattern analysis (loops/whorls/arches), confidence scores, and model evaluation metrics—all intgrtiy^iptpr^qe. Tlje systppi enhffleeg user^xpgriynec hv prowtim
Disclaimer: Curated by HT Syndication.
