Artificial intelligence

Scientific journal

ISSN 2710-1673

ONLINE: ISSN 2710-1681

Search by:

Year of publication
Author name
Paper title

Method of emotional condition recognizing by voice using phrases models

Klymenko M.1
1 Institute of artificial intelligence problems of MES and NAS of Ukraine

Full text (PDF)

UDC: 004.89:004.93
Publication Language: Ukrainian
Stuc. intelekt. 2017; 22(2):52-59

Abstract: In the article the method of emotional condition recognizing by voice is described, in which models of emotional conditions are formed by a set of acoustic, prosodic and extra-linguistic characteristics. The grouping of characteristics by their placement in typical fragments of intonational constructions is proposed. Numeric researches have shown that this approach made it possible to increase the likelihood of emotional recognition of emotional states compared to the results without the use of grouped characteristics.

Keywords: speaker model, acoustic characteristics of emotions, the Gaussian mixture model.

References:

  1. Klimenko N.S. Issledovanie effektivnosti bustinga v zadache tekstonezavisimoy identifikatsii diktora /N.S. Klimenko, I.G. Gerasimov // Shtuchniy Intelekt. – 2014. – №4 (66). – C. 191-201.
  2. Lobanov B.M. Kompternyiy sintez i klonirovanie rechi / B.M. Lobanov, L.I. Tsirulnik. – Minsk: "Belorusskayanauka". – 2008. – 316 s.
  3. Figueiredo M.A.T. Unsupervised Learning of Finite Mixture Models / M.A.T. Figueiredo, A.K. Jain // IEEETransactions on Pattern Analysis and Machine Intelligence. – 2002. – №24 (3). – P. 381–396.
  4. Psihologiya emotsiy / Izard K.E. Perev. s angl. - SPb: Izdatelstvo "Piter", 1999. - 464 s.

View full text (PDF)