An Enhanced Speech Emotion Recognition System using RNN compared with CNN

The aim of the work is to design a secure and accurate Speech Emotion Recognition (SER) system using a RNN model to classify emotions from speech data. The system’s predictive accuracy is evaluated and compared with the performance of a standalone CNN-based model. The suggested RNN model achieves an accuracy of 96.96%, significantly better than the CNN model’s accuracy of 93.93%. Group 1 has a CNN-based Speech Emotion Recognition system that was evaluated using ten models of different accuracy. Group 2 includes an advanced RNN model with convolutional and recurrent layers to maximize predicting accuracy. The CNN-based model achieves an accuracy range of 90.20% to 93.93%, while the RNN model demonstrates superior performance with an accuracy range of 93.80% to 96.96% under similar conditions. Testing parameters include a maximum difference of 2.50 and a minimum of 1.20, with a step size of 0.10 and a confidence level of p < 0.05. In this work, it is seen that the RNN model has significantly higher accuracy than the CNN model.