Development an Accurate Neural Network for Coin Recognition

This paper presents recognition system for coins of the USSR that is based on a convolutional neural network and database that stores images of coins and information about each coin. In addition, two algorithms for processing the incoming image were investigated, which occurs in four stages: cropping the background of the image, finding the center of the coin by rotating it, reducing the size of the coin image to increase the recognition speed of the neural network and find the maximum number of collinear vectors.