Penerapan Metode MFCC dan LSTM untuk Pengenalan Emosi melalui Suara

Abstract

Speech recognition falls under the field of computational linguistics. This includes identification, recognition, and translation of speech detected into text by a computer. This research uses Mel Frequency Cepstral Coefficients (MFCC) and Recurrent Neural Networks (RNN) techniques as a form of Artificial Neural Networks (ANN) architecture. The main purpose of this research is to use speech recognition techniques to detect and identify various emotional sounds in a person. The MFCC process will convert the voice signal into several vectors that help for the speech recognition process in this study.