Інформаційна технологія розпізнавання вимовлених літер на основі спайкінгової нейронної мережі

Abstract

An information technology for recognizing spoken letters based on a spiking neural network is proposed. The architecture of the chosen type of network was developed. A spoken letters recognition program written in the Python programming language in the PyCharm environment using the PySNN framework was designed. The developed program has a higher accuracy (61.7%) than the analog program (53.3%), which means that the accuracy of recognizing spoken letters has improved by 8.4%.