Blood Group Detection Using Image Processing and Deep Learning
Varun Kumar E , B, H Yashvanth, J Srinath · International Journal for Research in Applied Science and Engineering Technology · 2025
Abstract: Blood group identification is a critical pre-transfusion procedure that ensures patient safety. Current manual methods are time-consuming (5-10 minutes per test) and susceptible to human interpretation errors (estimated 2-5% inaccuracy). This paper presents an automated system that combines classical computer vision techniques (Scale-Invariant Feature Transform - SIFT and Oriented FAST and Rotated BRIEF - ORB) with a Convolutional Neural Network (CNN) to classify blood groups from microscopic slide images. The proposed architecture achieves 94.2% accuracy on a dataset of 1,200 samples across 8 blood types (A+, A-, B+, B-, AB+, AB-, O+, O-), reducing processing time to under 60 seconds. The system implements a novel dualvalidation approach where SIFT/ORB feature matching provides initial classification, followed by CNN verification. Clinical trials at [Hospital Name] showed 98% concordance with standard tube methods, demonstrating viability for emergency and resource-limited settings.