Analysis of Speech Signal Processing Methods in Speech Recognition Systems
Nomaz MirzaevTashkent University of Information Technologies named after Muhammad al-Khwarizmi,Tashkent,UzbekistanJohongir UrinboevDigital technologies and artificial intelligence research institute,Tashkent,UzbekistanNuraddin GafforovDigital technologies and artificial intelligence research institute,Tashkent,UzbekistanMavluda NugmanovaDigital technologies and artificial intelligence research institute,Tashkent,Uzbekistan
2025
ABI
Abstract
This work provides an overview of modern methods for processing speech signals used in speech recognition systems. Processing methods in the time, frequency, and frequency-time domains have been analyzed. Key approaches such as Fourier transform, wavelet transform, cepstral analysis, linear prediction methods, correlation analysis, using neural network, hidden Markov models, and dynamic time warping are examined. The conducted research allows for an assessment of the capabilities of existing methods, their effectiveness, as well as promising directions for the implementation of new mathematical apparatuses.
Topics
Identifiers
Citations and references
Cited by 08 references
Metrics — AkademScholar · Coming soon