UNIV. NEWS. NORTH-CAUCAS. REG. TECHNICAL SCIENCES SERIES. 2016; 1: 3-9
http://dx.doi.org/10.17213/0321-2653-2016-1-3-9
ALGORITHM FOR DYNAMIC SCALING ADAPTATION PATTERN FUNCTIONS
Gavrikov Mikhail Mikhailovich – Candidate of Technical Sciences, assistant professor, department «Software Engineering», Platov South-Russian State Polytechnic University (NPI), Novocherkassk, Russia. Ph. (8635)25-52-95. E-mail: gmm1000@yandex.ru
Sinetsky Roman Mikhailovich – Candidate of Technical Sciences, department «Software Engineering», Platov South-Russian State Polytechnic University (NPI), Novocherkassk, Russia. Ph. (8635)25-52-95. E-mail: rmsin@srspu.ru
Abstract
An algorithm for scale adaptation of the functions of the prototype representing classes of images performs identification of parameters of nonlinear deformation of scale contained in the functions representing the observed realization of images, and the subsequent transformation of the original function of the prototype using the previously proposed model. The algorithm works on the class of images of regular phase processes with the property quasi-similar shapes. Two models of scale nonlinear transformations - symmetrical and asymmetrical – are proposed. The experimental validation of the algorithm on images of prototype fragments voice and ECG. The examples and experimental data to support the effectiveness of the algorithm.
Keywords: computer modeling; regular phase process; signal analysis; speech recognition; electrophysiological signal processing
Full text: [in elibrary.ru]
References
1. Gavrikov M.M. Strukturnaya approksimatsiya i raspoznavanie odnomernykh vremennykh obrazov. Kontseptsiya i primeneniya [Structural approximation and recognition of the time images. The concept and application]. Izv. vuzov. Elektromekhanika= Russian Electromechanics, 2003, no. 6, pp. 52-60. [In Russ.]
2. Gavrikov M.M., Ivanchenko A.N. Konstruirovanie algoritmov strukturnoi approksimatsii signalov [Construction of signals structural approximation algorithms]. Izv. vuzov. Elektromekhanika= Russian Electromechanics, 1995, no. 1-2. pp. 104-112. [In Russ.]
3. Gavrikov M.M., Sinetskii R.M. Algoritmy modelirovaniya signalov i spektral'nykh funktsii s «pul'siruyushchimi» masshtabnymi iskazheniyami [Algorithms for modeling signals and spectral functions of a "pulsating" scale distortion]. Izvestiya vuzov. Severo-Kavkazskii region. Tekhnicheskie nauki, 2013, no. 3, pp.3-9. [In Russ.]
4. Gavrikov M.M., Sinetskiy R.M. Temporal structural images segmentation algorithms in the processing of speech signals [Algoritmy segmentatsii strukturnykh vremennykh obrazov i ikh primenenie v obrabotke rechevykh signalov]. Izvestiya vuzov. Severo-Kavkazskii region. Tekhnicheskie nauki, 2010, no. 1, pp.18-24. [In Russ.]
5. Fainzilberg L.S. Informatsionnye tekhnologii obrabotki signalov slozhnoi formy. Teoriya i praktika [Information technology of complex shape signals processing. Theoryandpractice]. Ukraina, Kiev, «Naukova dumka», 2008, 336 pp.
6. Gavrikov M.M., Sinetskiy R.M. Algoritmicheskaya i chislennaya realizatsiya strukturno-approksimatsionnogo metoda raspoznavaniya rechevykh obrazov [Algorithmic and numerical implementation of structural-approximation method in speech recognition]. Izv. vuzov. Elektromekhanika= Russian Electromechanics, 2007, no. 2, pp. 52-59. [In Russ.]
7. Allen J.B. Short-term spectral analysis, synthesis, and modification by discrete Fourier transform // IEEE Trans. on Acoustics, Speech, Signal Processing. 1997. Vol. ASSP-25. № 3. pp. 235–238.
8. Gavrikov M.M., Sinetskiy R.M. Tekhnologiya sinteza strukturno-approksimatsionnykh etalonov rechevykh obrazov v komandno-rechevykh interpretatorakh [The technology of synthesis of structural approximation patterns of speech images in voice-command interpreters]. Izv. vuzov. Elektromekhanika= Russian Electromechanics, 2005, no. 1, pp. 40-46. [In Russ.]
9. Zlatoustova L.V., Potapova R.K., Trunin-Donskoy V.N. Obshchaya i prikladnaya fonetika [General and applied phonetics]. Moscow, MGU Publ., 1986, 304 pp.
10. Gavrikov M.M., Ivanchenko A.N. Metod strukturnoi approksimatsii v obrabotke signalov i eksperimental'nykh krivykh [The method of structural approximation in processing of signals and experimental curves]. Izv. vuzov. Elektromekhanika= Russian Electromechanics, 1991, no. 5, pp. 67-79. [In Russ.]
11. Fainzilberg L.S. Heart functional state diagnostic using pattern recognition of phase space ECG-images // Proc. 6th European Congress on Intelligent Techniques and Soft Computing (EUFIT ’98). Aachen (Germany). 1998. № B-27. pp. 1878–1882.