350 rub
Journal Neurocomputers №10 for 2012 г.
Article in number:
Implementation of spoken words neural network recognition on a mobile robotic platform based on signal processors
Keywords: isolatedwordrecognition neural networks robotics mel frequency cepstral coefficients speaker-independent speech recognition noise compensation robotic platform signal processors
Authors:
M.A. Yakovlev, P.V. Skribtsov, M.A. Chervonenkis, S.N. Zagoruyko
Abstract:
Addressed in this paper are the principals of word recognition system implementation on a mobile robotic platform. Two classification algorithms used for recognition were studied: an algorithm based on k-nearest neighbors method and neural network algorithm. Article describes common approaches to implementation as well as peculiarities of design which are required for a system to operate on a robotic platform.
Pages: 21-25
References
  1. Официальный сайт проекта CMUSphinxhttp://cmusphinx.sourceforge.net/
  2. Официальный сайт проекта Juliushttp://julius.sourceforge.jp/en_index.php/
  3. Lamel, L. F., Rabiner, L. R., Rosenberg, A. E., Wilpong, J. G., An improved endpoint detection for isolated word recognition // IEEE Transactions on acoustics, speech, and signal processing. AUGUST 1981. V. ASSP-29. № 4.
  4. Описание стандарта ETSI201 108http://www.etsi.org/deliver/etsi_es/201100_201199/201108/01.01.03_60/es_201108v010103p.pdf
  5. Ratanamahatana, C. A.,Keogh, E., Three Myths about Dynamic Time Warping Data Mining // Department of Computer Science and Engineering University of California, Riverside.
  6. Choi, E. H. C., Noise Robust Front-end for ASR using Spectral Subtraction, Spectral Flooring and Cumulative Distribution Mapping - Multimodal User Interaction National ICT Australia.
  7. Vinyals, O., Ravuri, S. V., Povey, D., Revisiting recurrent neural networks for robust ASR // International Computer Science Institute, Berkeley. CA. USA EECS Department, University of California. Berkeley. Berkeley. CA, USAMicrosoft Research, Redmond. WA.USA.