Hybrid Approach for Human Emotion Recognition from Speech

摘要

Even though emotions don’t have much to do with the content of the speech, it has a major impact on human communication by providing much more positive feedback. Therefore, Speech emotion recognition (SER) and multimodal emotion recognition systems have been a hot area of research owing to their range of applications in several domains, such as social robots, virtual reality, and human-machine interaction applications. This paper compares two models by choosing multi-dimension CNN models and features for SER on the RAVDESS dataset.