Abstract—This paper presents a method for speech recognition using hybrid of dynamic vocabulary through dynamic grammar and parallel recognition events in runtime environment. The key idea is to use different recognition operations in such a way that speech can be recognized accurately. The method makes the recognition of the speech independent of old static vocabulary through a dynamic grammar loading process. The proposed approach is divided into different tasks which are Dynamic grammar upload and Parallel listening events for recognition process. This algorithm can quickly and correctly detect the speech. Under the experimental database which was taken from real speech inputs, 247 out of 250 words were successfully recognized. The average accuracy of speech recognition is, therefore, 85.0%.
Index Terms—Dynamic grammar, parallel listening events, speech recognition, parallel recognition events, runtime environment.
Muhammad Sharif, Syed Zia Uddin Bukhari, Waseem Khan and Mudassar Raza are with the Department of Computer Sciences, COMSATS Institute of Information Technology, Wah Cantt., Pakistan (e-mail: muhammadsharifmalik@yahoo.com, muhammad.wasim1@gmail.com, mudassarkazmi@yahoo.com).
Aman Ullah Khan is with the Department of Computer Science and Engineering, Air University Multan Campus, Pakistan (e-mail: auk_pk@yahoo.com).
[PDF]
Cite:Muhammad Sharif, Syed Zia Uddin Bukhari, Aman Ullah Khan, Waseem Khan, and Mudassar Raza, "Speech Recognition Using Dynamic Grammar and Parallel Listening Events," International Journal of Future Computer and Communication vol. 3, no. 2, pp. 94-97, 2014.