Determined key phrases search and recognition in continuous speech

Authors: Levinskiy A.T., Bykovskiy M.K.
Published in issue: #11(28)/2018
DOI: 10.18698/2541-8009-2018-11-399
Category: Informatics, Computer Engineering and Control | Chapter: Methods and Systems of Information Protection, Information Security
Keywords: recognition, error, interface, testing, Levenshtein method, w-shingling, comparison, function
Published: 12.11.2018

The problem of recognition of Russian speech and of lack of software to meet the needs of consumers is considered. Speech recognition systems are computing systems that allow to select continuous speech of the speaker from the general flow of sounds and noises. These systems are based on speech recognition technology that converts spoken words from sounds to digital clock signals using a recognition process. As a solution, software is proposed for speech recognition and searching for the required information in the received text. The subject area is analyzed, interested parties and their objectives are identified, on the basis of which the system usage scenarios is compiled and processed into basic functions. A review of existing software and algorithms is performed. Recognition systems are compared.

References

[1] Galunov V.I., Solov’yev A.N. Modern problems in field of speech recognition. Informatsionnye tekhnologii i vychislitel’nye sistemy, 2004, no. 2, pp. 42–45.

[2] Zapryagaev S.A., Konovalov A.Yu. Speech signals recognition. Vestnik voronezhskogo gosudarstvennogo universiteta. Ser. Sistemnyy analiz i informatsionnye tekhnologii [Proceedings of Voronezh State University. Series: Systems analysis and information technologies], 2009, no. 2, pp. 39–48.

[3] Tsentr rechevykh tekhnologiy (TsRT) [Center of speech technologies (TsRT)]. Available at: http://www.speechpro.ru/innovation/ (accessed 12 May 2017).

[4] Leont’yev A.A. Yazyk, rech’, rechevaya deyatel’nost’ [Language, speech, speech activity]. Moscow, URSS publ., 2007, 212 p.

[5] Gusev M.N. Metody i modeli raspoznavaniya russkoy rechi v informatsionnykh sistemakh [Methods and models of Russian speech recognition in information systems]. Sankt-Petersburg, SPbSUT publ., 2014, 378 p.

[6] Rechevye tekhnologii SpeechKit. Cloud [SpeechKit speech technologies. Cloud]. Available at: https://tech.yandex.ru/speechkit/cloud/ (accessed 24 May 2017).

[7] Pal’shina D.A. Speech rate as a reason for reduced forms of Russian words in everyday communication. Vestnik Permskogo universiteta. Rossiyskaya i zarubezhnaya filologiya [Perm University Herald. Russian and Foreign Philology], 2013, no. 2(22), pp. 18–24.

[8] Mazurenko I.L. Komp’yuternye sistemy raspoznavaniya rechi [Computer systems of speech recognition]. Moscow, URSS publ., 2017, 56 p.

[9] Zadacha raspoznavaniya rechi poka ne reshena [Speech recognition problem is not solved yet]. Available at: https://habr.com/post/408017/ (accessed 23 March 2018).

[10] “Cloud, raspoznavanie rechi, tekhnologii Yandeksa” [V Internete]. Available at: https://tech.yandex.ru/speechkit/cloud/doc/guide/ (accessed 11 June 2018).