MEL Frequency Features Driven Deep Learning for Speech Emotion Detection: A Novel Model

Speech Emotion Recognition (SER) is a critical area of research in Human-Computer Interaction (HCI), aiming to develop systems capable of identifying and responding to human emotions conveyed through speech. This paper focuses on training a Deep Neural Network (DNN) model using Mel-Frequency Cepstral Coefficients (MFCCs) for feature extraction, a widely used and effective technique in this domain. To address the complexity of speech data, we utilized a deep learning algorithm, specifically employing the Long Short-Term Memory (LSTM) network, in conjunction with MFCC features.