Generative Models for Automatic Speech Recognition, Understanding and Synthesis
A generalised generative model for automatic dictation and spoken translation machine is proposed. The model is based on both the generative grammar hierarchy for speech model signals composition and the comparison of them with a signal to be recognised. The details of problem solving at all levels of the speech processing hierarchy are discussed.
KeywordsSpeech Recognition Speech Signal Canonical Form Model Signal Automatic Speech Recognition
Unable to display preview. Download preview PDF.
- 1.T.K. Vintsiuk. Speech Word Recognition by Dynamic Programming. — Kibernetika, 1968, No 1, pp 81–88.Google Scholar
- 2.T.K. Vintsiuk. Element-Wise Recognition of Continuous Speech Composed of Words from Specified Dictionary. — Kibernetika, 1971, No 2, pp 133–143.Google Scholar
- 3.T.K. Vintsiuk. Phoneme Recognition of Connected Speech. — Avtomatyka, 1972, No. 6, pp 40-49; 1973, No 1, pp 63–72.Google Scholar
- 4.T.K. Vintsiuk. Speech Recognition and Understanding. — Kibernetika, 1982, No. 5, pp. 101–111.Google Scholar
- 5.T.K. Vintsiuk. Analysis, Recognition and Understanding of Speech Signals. — Kiev: Naukova Dumka, 1987, 264 p (in Russian).Google Scholar
- 6.T.K. Vintsiuk. HCDP-Technique for Automatic Analysis, Recognition and Understanding of Speech Signals. — Proc. First Intern. Conf. on Information Technology for Image Analysis and Pattern Recognition, L’viv, 1990, Vol 1, pp 108–112.Google Scholar
- 7.Taras Vintsiuk. Intellectual Speech Technologies and Systems in Ukraine. — In: Signal/Image Processing and Pattern Recognition, Kyjiv, 1996, pp 117–120 (in Ukrainian).Google Scholar
- 8.T.K. Vintsiuk. Two Approaches to Create a Dictation/Translation Machine. - Proc. of the 2nd Intern. Workshop “Speech and Computer”, Cluj-Napoca, 1997, pp 1–6.Google Scholar