Emotion recognition from speech signal using epoch parameters

Abstract

This paper proposes epoch parameters extracted from LP (Linear Prediction) residual and zero frequency filtered speech signal for recognising the emotions present in speech. Instant of glottal closure within pitch period of LP residual is known as an 'epoch'. The significant excitation of vocal tract usually takes place at the instant of glottal closure. In this paper the epoch parameters namely strength of epoch, instantaneous frequency, sharpness of epochs, slope of strength of epochs are used as features for classification of emotions. These features are extracted from the glottal closure region of LP residual. For analysing emotion recognition, using the proposed epoch parameters, actor recorded Telugu database (IITKGP-Simulated Emotion Speech Corpus) and Berlin emotional database are used. In the study we have considered six emotions namely anger, disgust, fear, happy, neutral and sadness. Gaussian mixture models and support vector machines are used for developing the models. Average emotion recognition of 61% and 58% is observed respectively for the above models.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 51 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.