Title :
Formants Based Analysis for Speech Recognition
Author :
Ali, Ahmad ; Bhatti, Safiullah ; Mian, Muhammad Sleem
Author_Institution :
Dept. of Comput. Sci. & Software Eng., Univ. of Eng. & Technol., Lahore
Abstract :
Automatic speech recognition (ASR) is one of the most developing fields of the modern science having a wide range of applications. The purpose of this paper is to introduce statistical methods for speech recognition by extracting formants of the speech and analyzing their behavior. It is a novel way to speech recognition. The whole analysis for speech recognition is based upon first five formants of the speech. The method has been tested for Urdu speech, and result obtained is of high accuracy. There are 80 checkpoints, as a result of the algorithm (method), before Urdu speech is recognized. The method can be used for any other language as well for speech recognition
Keywords :
speech recognition; statistical analysis; Urdu speech; automatic speech recognition analysis; formant based analysis; speech formant extraction; statistical method; Acoustical engineering; Automatic speech recognition; Frequency; Humans; Shape; Software engineering; Speech analysis; Speech recognition; Statistical analysis; Tongue;
Conference_Titel :
Engineering of Intelligent Systems, 2006 IEEE International Conference on
Conference_Location :
Islamabad
Print_ISBN :
1-4244-0456-8
DOI :
10.1109/ICEIS.2006.1703179
