Determinação do valor total de moedas em imagens digitais

Abstract

A common problem encountered by people with vision problems is the count and recognition of coins. To facilitate this task this paper proposes an algorithm to do this automatically. To enable detection and automatic coin counting, it is important to perform a count with a low error rate. One way to accomplish this count is to apply a classification algorithm. In this paper we test if this is a good approach. Conclude that it is possible to obtain an accuracy rate above 90% used decision trees in one of the proposed models.