Extracting multi band approach of acoustic vectors extractors:Usign HMM Classifier
Author(s):
LAMKADAM Abdelmajid , Faculty of Science, P.O Box 1796, Dhar El Mehraz, University of Sidi Mohamed Ben Abdellah, Fez, Morocco; KARIM Mohamed, University of Sidi Mohamed Ben Abdellah, Fez, Morocco
Keywords:
Acoustic Vectors Extraction, ACP, DFE, Diagonalization, Discrimination, HMM, LDA, LPC, MFCC, PLP, RASTA, Recognition, Recombination
Abstract:
Currently, speech recognition not yet optimized, it still has difficulties, in particular when it comes to the recognition in the Arabic language. This complexity is due to the ambiguities contained in the speech signal, and the various obstacles encountered during the processing of the signal. Indeed, a comparative study of extraction methods (LPC, MFCC, PLP, and RASTA) is made, and a review of existing combinations is carried out, to assess the rate of recognition of authentication systems. Our goal is reserved to develop an extractor’s recombination using the HMM classifier, with the aim of reaching a full treatment with a considerable amount of extracted vectors acoustic, in order to improvement the recognition rate of Arabic numerals.
Other Details:
Manuscript Id | : | IJSTEV3I2095
|
Published in | : | Volume : 3, Issue : 2
|
Publication Date | : | 01/09/2016
|
Page(s) | : | 305-310
|
Download Article