Development and Experimental Validation of a NonInvasive Blood Group Detection System

Abstract

An individual's blood group consists of red blood cell antigens whose composition is determined by protein presence, antigen structure, and gene series. Persons aged above six months have significant anti-A and/or anti-B in their serum. During transplantation and transfusion, ABO blood group identification is the most essential factor. The conventional method involves drawing blood samples from patients, and the blood group is determined based on the antigen-antibody reaction. This method consists of adding chemical reagents. However, this requires time of operation, and throughput analysis is high, and the process is also challenging to interpret. Accurate and rapid identification of blood groups is therefore crucial in various medical fields, including blood transfusions, organ transplants, and prenatal care. Traditional methods for blood typing often require extensive laboratory equipment and trained personnel, leading to delays and potential errors in critical situations. This research focuses on developing a non-invasive, compact, and user-friendly device capable of determining blood groups quickly without invasively collecting patient’s blood samples and using reagents. The system learns from a database of annotated blood samples by employing machine learning algorithms, enhancing its accuracy and reliability over time. A noninvasive blood group detection system was verified experimentally on a laboratory prototype, achieving an accuracy of 95.9% in identifying blood groups and rhesus factors. Furthermore, a comparative analysis was conducted between the proposed system and existing counterparts. This analysis demonstrated that the proposed system outperforms others in accuracy, indicating the rhesus factor