Approaches in Pattern Recognition and Classification Based on Coin Images

Abstract

This paper discusses the various texture-based and vector-based approaches to classify coins.These two types of approaches are common for software-based coin sorting systems. Many researchers have applied algorithms known in artificial intelligence research for feature extraction, selection and classification of coin images. However, without a common benchmark, it is difficult to assess the accuracy, robustness and efficiency of the coin sorting systems across the different approaches. It is proposed that the use of a standard benchmark image databank such as the CIS Benchmark will allow a more objective and accurate comparison of the performance of these coin classification approaches.