Optimal parameters selected for automatic recognition of spoken Amazigh digits and letters using Hidden Markov Model Too
- PDF / 1,894,235 Bytes
- 11 Pages / 595.276 x 790.866 pts Page_size
- 78 Downloads / 211 Views
Optimal parameters selected for automatic recognition of spoken Amazigh digits and letters using Hidden Markov Model Toolkit Safâa El Ouahabi1 · Mohamed Atounti1 · Mohamed Bellouki1 Received: 17 February 2020 / Accepted: 15 October 2020 © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract In this paper, we present our Amazigh automatic speech recognition system. Its realization is constructed with contextindependent phonetic Hidden Markov Models. Many choices are made on this system, such as the number of states of the models, the type of emission probability densities associated with the states, and the representation of the signal by cepstral coefficients. The results of recognition of our system place it at a level of height performance comparable to that achieved by Markovian automatic speech recognition systems. Our system is designed to recognize 43 distinct isolated Amazigh words (33 letters and 10 digits). The recognition rate is then calculated for each digit and letter. The overall accuracy and word recognition rate for the whole database achieved 91.31% after extensive testing and change of the recognition parameters. The results obtained in this work are improved in association with our previous work concerning Amazigh spoken digits and letters automatic speech recognition, using Hidden Markov Model Toolkit. Keywords Amazigh isolated words · Automatic speech recognition system (ASR) · Hidden Markov Model (HMM) · Hidden Markov Model Toolkit (HTK) · Gaussian mixture modelling (GMM) · Mel-frequency cepstral coefficients (MFCC)
1 Introduction The automatic speech recognition field has considerably progressed and revealed innovative algorithms and techniques for the statistics treatment of speech. Using natural language in the Human/machine interaction places this area at the center of interest of researchers in several domains such as signal processing, linguistics researches, artificial intelligence, and computer science. Unfortunately, despite the incredible progress of techniques and algorithms, automatic speech recognition remains in the process of research because the results achieved are still so far from being ideal, and the applications remain strongly dependent on the target language. ASR is located and applied in several fields namely the intersection of signal processing and language * Mohamed Atounti [email protected] Safâa El Ouahabi [email protected] Mohamed Bellouki [email protected] 1
Laboratory of Applied Mathematics and Information System, Polydisciplinary Faculty of Nador, Nador, Morocco
processing. It is, therefore, a research topic that leads the computer researchers to develop speech and language signal processing algorithms that satisfy the lexicological, syntactic, and semantic principles of the language developed. These algorithms mostly use speech as pertinent information to produce applications that can effectively process natural speech. In general, the performance of a speech recognition system depends mainly on the language studied,
Data Loading...
