Functional simulation of human blood identification device using feed-forward artificial neural network for FPGA implementation

The identification of human blood type still requires a fast and accurate device considering the number of blood samples that need to be distributed and transfused immediately. In this study we propose a hardware implementation of human blood type identification devices using feedforward neural network algorithms on grayscale images of blood samples. The images to be used are 32×32 pixels, 48×48 pixels, 64×64, 80×80, and 96×96 pixels. The algorithm were implemented using VHSIC Hardware Description Language. With artifical neural network implemented on Xilinx FPGA Spartan 3S1000, the success rate of detection by grouping by the mean and median ratios of the number of `1' bits is more than 75%.