Survey on Non-Invasive Blood Group Detection Using Image Processing

Abstract

Detecting the blood group of people is necessary, before performing a transfusion of blood in an urgent situation or when checking a person’s blood group for donation. Presently, the tests are conducted manually by the lab personnel in the laboratory. This takes time and consequences in human error when determining blood type. This may result in loss of life or seriously affect people’s lives. To overcome this, we propose a non-invasive model using image processing and machine learning algorithms. Image processing is useful in various ways for researchers to succeed in their goals, especially in medical fields. This system displays the corresponding blood type that is A+, A-, B+, B-, AB+, AB-, O+, and O- in a short time. Image matching techniques like ORB are used to extract features from an image. It involves signal extraction processes, spectroscopic information, feature calculation processes, and Photoplethysmographic (PPG) signal. The features extracted are given to a machine learning model to detect the blood type of a person in a short period. This method does not require mixing any antigens like A, B, and O with blood samples. This method does not use any blood samples, does not require any chemicals, and reduces human error.