A Novel Approach for ABO Blood Group Prediction using Fingerprint through Optimized Convolutional Neural Network

Abstract

The fingerprints of humans hold many potentials and having plenty of unique characteristics. It is one of the primary diagnostic tools used from accent's era because of their distinctive identity. It opens a lot of possibilities for human science research, by analyzing fingerprints, some researchers have tried to predict an individual's gender or age. In the 20th century, fingerprint identification and analysis have become commonplace also it has evolved into a crucial component of forensic at crime scenes. Similarly, like a fingerprint, the blood group is also unique for each individual. This study focuses on ABO and Rh systems, which are among the most prominent blood grouping methods. This paper proposed an optimized Convolutional Neural Network (CNN) which is designed as an extension of an AlexNet, that correlates the fingerprint patterns or different features of the fingerprint with the blood group of an individual. Researchers have only attempted to connect fingerprint patterns with blood types prior to this proposed method. The result and performance of proposed CNN framework is compared with three different CNN variations like LeNet 5, ZFNet, and AlexNet. The design of proposed CNN used for the prediction of the blood group having noticeable performance with 95.27 % accuracy rate.