An Intelligent Emotion Recognition System based on Speech Terminologies using Artificial Intelligence Assisted Learning Scheme

From HCI to mental health monitoring, emotion identification from voice signals is an important and vital task. We provide a novel method for reliable emotion identification from audio data called Artificial Intelligence Assisted Learning Scheme (AIALS). AIALS combines the architectures of Long Short-Term Memory (LSTM) and Convolutional Neural Network (CNN). The CREMA-D dataset, which stands for Crowd-sourced Emotional Multimodal Actors Dataset, is used in our work to train and assess the suggested system. The dataset is rich in authentic emotional speech examples. Feature extraction and emotion classification are the two primary components of the AIALS framework that has been suggested. During the feature extraction stage, pertinent acoustic information is retrieved from raw speech signals by computing Mel-frequency Cepstral coefficients (MFCCs). Subsequently, spectrograms are generated to extract spatial features using the CNN architecture. In the emotion classification stage, the extracted features are fed into the hybrid LSTM-CNN model, which effectively learns temporal dependencies and spatial patterns from the input data. Our experimental results demonstrate the effectiveness of the proposed AIALS framework in accurately recognizing emotions from speech signals. By leveraging the complementary strengths of LSTM and CNN architectures, our model achieves an impressive accuracy of 98% on the CREMA-D dataset, outperforming existing state-of-the-art approaches. Furthermore, the proposed system exhibits robustness across different emotional categories, showcasing its generalization capability.