Translate this page:
Please select your language to translate the article


You can just close the window to don't translate
Library
Your profile

Back to contents

Software systems and computational methods
Reference:

Sidorov K.V., Filatova N.N. Automatic recognition of human emotions based on the reconstruction of the speech samples attractors

Abstract: The article reviews the methods of automatic pattern recognition of the speech signals recorded at the moments when announ cers proved positive emotions from speech samples on the same subjects recorded in a testees’ neutral state. This article investigates the abilities of nonlinear dynamics methods for evaluation of informative indicators of emotional state. Studies were performed on the basis of analyzing the reconstruction of attractors of the speech signal. The authors analyzed different ways of selecting optimal parameter values for the reconstruction of the attractor (the time delay between the elements of the time series and embedding dimension). Authors proposed the new quan titative criteria for classifying samples of the speech signal of a person experiencing emo tions based on the estimates of the maximum vector reconstruction of the attractor in four quadrants. The research was based on frag ments of the Russian-language database (Tver). A model of emotional body language, which consists of a database of two levels (phrases and phonemes) was formed and served as a basis for evaluation of the efÞ cie n cy of the developed software module of the automatic recognition of human emotions.


Keywords:

Software, speech, speech signal, emotional state, time series, emotion, emotion recognition, nonlinear dynamics, attractor reconstruction, classification


This article can be downloaded freely in PDF format for reading. Download article


References
1. Davydov A.G., Kiselev V.V., Kochetkov D.S. Klassifikatsiya emotsional'nogo sostoyaniya diktora po golosu: problemy i resheniya // Trudy mezhdunarodnoy konferentsii «Dialog 2011». – M.: RGTU, 2011. – S. 178–185.
2. Luk'yanitsa A.A., Shishkin A.G. Avtomaticheskoe opredelenie izmeneniy emotsional'nogo sostoyaniya po rechevomu signalu // Rechevye tekhnologii. – M.: Narodnoe obrazovanie, 2009. – ¹3. – S. 60–76.
3. Golubinskiy A.N. Vyyavlenie emotsional'nogo sostoyaniya cheloveka po rechevomu signalu na osnove veyvlet-analiza // Vestnik Voronezhskogo instituta MVD Rossii. – 2011. – ¹3. – S. 144–153.
4. Sidorov K.V., Filatova N.N. Analiz priznakov emotsional'no okrashennoy rechi // Vestnik Tverskogo gosudarstvennogo tekhnicheskogo universiteta. – Vyp. 20. – Tver', 2012. – S. 26–31.
5. Starchenko I.B., Perervenko Yu.S., Borisova O.S., Momot T.V. Metody nelineynoy dinamiki dlya biomeditsinskikh prilozheniy // Izvestiya YuFU. Tekhnicheskie nauki. Tematicheskiy vypusk «Meditsinskie informatsionnye sistemy». – Taganrog: TTI YuFU, 2010. – ¹ 9 (110). – S. 42–51.
6. Makarova V., Petrushin V.A. RUSLANA: a database of russian emotional utterances // ICSLP, 2002. – pp. 2041–2044.
7. http:// www.harpia.ru/rec/.
8. Kalyuzhnyy M.V. Sistema reabilitatsii slabovidyashchikh na osnove nastraivaemoy segmentarnoy modeli sinteziruemoy rechi: avtoref. dis. …kand. tekh. nauk: 05.11.17. – SPb., 2009. – 18 s.
9. Sidorov K.V., Filatova N.N. Algoritm avtomaticheskoy generatsii rechevykh ob'ektov // Sbornik materialov I Mezhdunarodnoy nauchn.-prakt. konf. «Tekhnicheskie nauki-osnova sovremennoy innovatsionnoy sistemy». – Chast' 1. – Yoshkar-Ola, 2012. – S. 118–120.
10. Takens F. Detecting strange attractors in turbulence // Dynamical Systems and Turbulence. – Heidelberg: Springer-Verlag, 1981. – pp. 366–381.
11. Gorshkov V.A., Kasatkin S.A. Identifikatsiya vremennykh ryadov aviatsionnykh sobytiy metodami i algoritmami nelineynoy dinamiki. – M.: Blank Dizayn, 2008. – 208 s.
12. Malinetskiy G.G., Potapov A.B. Sovremennye problemy nelineynoy dinamiki. – M.: Editorial URSS, 2000. – 336 s.
13. Mekler A.A. Primenenie apparata nelineynogo analiza dinamicheskikh sistem dlya obrabotki signalov EEG // Vestnik novykh meditsinskikh tekhnologiy. – 2007. – T. KhIV, ¹ 1. – S. 73–76.
14. Kennel M.B., Brown R., Abarbanel I. Determining embedding dimension for phase-space reconstruction using a geometrical construction // Phys. Rev. A, 45, 3403, 1992.
15. Hegger R. et al. Practical Implementation of Nonlinear Time Series Methods. In: The TISEAN package, CHAOS 9, 413. – 1999. – http://www.mpipks-dresden.mpg.de/~tisean/