24-04-2017, 10:28 AM
Emotional recognition of speech has emerged as an important area of research in the recent past. In this sense, the review of existing work on emotional speech processing is useful for conducting further research. In this paper, the recent literature on the recognition of speech emotion has been presented considering the topics related to the corpus of emotional speech, the different types of speech characteristics and the models used for the recognition of the emotions of speech. In this paper, thirty-two representative databases of speech are reviewed, from the point of view of their language, number of speakers, number of emotions and purpose of the collection. Problems related to the emotional speech databases used in the recognition of emotional speech are also briefly discussed. Literature is presented on the different characteristics used in the task of recognition of the emotion of speech. The importance of choosing different classification models has been discussed together with the review.
Although the detection of emotions from speech is a relatively new field of research, it has many potential applications. In human-human or human-human interaction systems, emotion-recognition systems could provide users with improved services by adapting them to their emotions. In virtual worlds, recognizing emotions could help simulate a more realistic avatar interaction. The body of work on detecting emotion in speech is quite limited. At present, researchers are still debating which features influence the recognition of emotion in speech. There is also considerable uncertainty as to the best algorithm for classifying emotion, and what emotions to class together. In this project, we try to address these issues. There are a variety of temporal and spectral features that can be extracted from human speech. We used statistics related to tone, Mel Frequency Cepstral Coefficients (MFCCs) and Formants of speech as inputs to the classification algorithms. The precision of emotional recognition of these experiments allows us to explain which characteristics carry the most emotional information and why. It also allows us to develop criteria to classify the emotions together.