Optimized Image Processing Algorithms for Precise Blood Group Detection

Abstract

The need for a reliable and non-invasive method for blood group detection is increasingly recognized, particularly in regions with limited access to conventional medical testing. This paper presents a deep learning approach that utilizes fingerprint images to determine blood groups. A convolutional neural network architecture was implemented to analyze distinct patterns in fingerprint data, enabling the classification of eight blood groups. An initial model was developed to establish a baseline for classification, and its performance was evaluated through loss and accuracy metrics. Results indicate the potential for this fingerprint based method to serve as an accessible point of care solution for blood group determination, paving the way for further enhancements in model complexity and accuracy.