Title :
Detection Of Endpoints Of Isolated Words Using Slope Transformation
Author :
Marciniak, Tomasz ; Rochowniak, R. ; Dabrowski, Adrian
Author_Institution :
Poznan Univ. of Technol.
Abstract :
In this paper an algorithm for detection of endpoints of isolated words, which is a part of the automatic speech recognition (ASR) system, is presented. Our algorithm is based on the Legendre (slope) transformation. In order to analyze an influence of the proposed algorithm on the recognition effectiveness, a set of 10 Polish steering words was tested in two characteristic cases: recognition of original (unnoised) words and recognition of words noised at level of -3 dB. The achieved recognition rate was for original and for noised words equal to 98.48% and 97.84%, respectively
Keywords :
natural languages; speech recognition; transforms; ASR system; Legendre transformation; Polish steering words; automatic speech recognition; isolated word endpoints detection; slope transformation; Algorithm design and analysis; Automatic speech recognition; Automatic testing; Character recognition; Fault detection; Filtering; Hidden Markov models; Isolation technology; Noise level; Noise robustness;
Conference_Titel :
Mixed Design of Integrated Circuits and System, 2006. MIXDES 2006. Proceedings of the International Conference
Conference_Location :
Gdynia
Print_ISBN :
83-922632-2-7
DOI :
10.1109/MIXDES.2006.1706665
