EUROSPEECH '97
5th European Conference on Speech Communication and Technology

Rhodes, Greece
September 22-25, 1997


Empirical Comparison of Two Multilayer Perceptron-Based Keyword Speech Recognition Algorithms

Suhardi (1), Klaus Fellbaum (2)

(1) Institute for Telecommunication and Theoretical Electrical Engineering, Technical University of Berlin, Germany
(2) Communication Engineering, Brandenburg Technical University of Cottbus, Germany fellbaum@kt.tu-cottbus.de

In this paper, an empirical comparison of two multilayer perceptron (MLP)-based techniques for key- word speech recognition (wordspotting) is described. The techniques are the predictive neural model (PNM)-based wordspotting, in which the MLP is applied as a speech pattern predictor to compute a local distance between the acoustic vector and the phone model, and the hybrid HMM/MLP-based wordspotting, where the MLP is used as a state (phone) probability estimator given acoustic vectors. The comparison was performed with the same database. According to our experiments, the hybrid HMM/MLP-based technique excels the PNM-based techniques (~6.2 %).

Full Paper

Bibliographic reference.  Suhardi / Fellbaum, Klaus (1997): "Empirical comparison of two multilayer perceptron-based keyword speech recognition algorithms", In EUROSPEECH-1997, 2835-2838.