Blood Typing Platform using a Local Binary Pattern from Blood Sedimentation in a Microplate

Abstract

Blood donation plays a vital role in saving human life and also indicates a essential quality of local health system. In the practical of blood donation, blood grouping for ABO and Rh systems is required according to the international standards of the blood bank in hospitals. Which, relying on a technician to identify the results that interpret by the coagulation pattern of blood was mixed with the antibodies against the blood type. This research proposes a platform for grouping blood types by a specific binary model. The different coagulation patterns for each blood group were performed using the Local Binary Pattern (LBP) method consisting of eight coagulation characteristics of blood groups. The platform can recognize blood types in the ABO system and the Rh system at the same time. Data collection is performed with coagulation and non-sediment imaging of blood samples with antibodiesrecognized to the ABO system and the Rh system blood group. The data analysis on this platform is developed in a Python programming language and processed on a Raspberry Pi. In results, it can be noted that the analysis of blood groups to be accurate, fast and able to analyze multiple samples at the same time. In addition, the data collection would be generated and recorded onto the website of the blood donor information and also can be connected to the hospital database in the future. In the results, a system demonstration using photographs of blood donor testing of 120 donors who the donor blood has been mixed with antibodies. It was found that the proposed system was able to correctly identify all donor blood groups and takes an average of 2.4 people per second processing time.