Abstract
As a method of discriminating invisible part information, there is a hammering test for discriminating an object using a sound signal. In this paper, we deal with coin falling sounds as an example of such sound classification. In conventional coin discrimination studies, humans have been analyzing a single speech feature. However, it is difficult to accurately analyze many speech features by this method. It is necessary to accurately analyze large quantities of features. Combining machine learning with sound signals analysis seems to make it possible to analyze many feature quantities more accurately. For this reason, we analyzed sounds of popular five-yen and ten-yen coins falling into various materials using machine learning.
