DEEP LEARNING-BASED FINGERPRINT RECOGNITION FOR MEDICAL DIAGNOSIS: A STUDY ON BLOOD GROUP, GLUCOSE LEVEL, AND PLATELET COUNT DETECTION

This study introduces an innovative method for identifying blood group, glucose concentration, and platelet count through the analysis of fingerprint images using deep learning.A convolutional neural network (CNN) is implemented to derive distinctive features from fingerprints, which serve as predictors for the aforementioned medical indicators.The model is trained on an extensive set of annotated fingerprint images, showing promising performance with detection accuracies exceeding 95% for blood groups, 90% for glucose levels, and 85% for platelet counts.The proposed approach is non-invasive, economical, and efficient, holding substantial potential to transform diagnostic practices, particularly in underserved regions.