.

SPEECH RECOGNITION USING

NEURAL NETWORKS

ABSTRACT

Speech processing has been an active area for several decades with a
wide variety of applications ranging from communications to automatic reading
machines. The development of Speech Recognition Products is mainly based on
statistical techniques which work under very specific assumptions. The work presented
in this paper investigates the feasibility of an alternative approach for solving the
problem more efficiently using Artificial Intelligence (Neural Networks).

A speech recognizer system comprises of two distinct blocks, a Feature

Extractor and a Recognizer. The Feature Extractor block uses a standard LPC Cepstrum
coder, which translates the incoming speech into a trajectory in the LPC Cepstrum
feature space, followed by a Self Organizing Map, which tailors the outcome of the
coder in order to produce optimal trajectory representations of words in reduced
dimension feature spaces. Designs of the Recognizer blocks based on two different
approaches are compared. The performance of Multi-Layer Perceptrons, and Recurrent
Neural Networks based recognizers is tested.

We took the samples of the ten digits (0,1,2…) by the same speaker and
examined the accuracy rates based on two different approaches mentioned above.
Experimental results indicated that trajectories on such reduced dimension spaces can
provided reliable representations of spoken word, while reducing the training
complexity and the operation of the Recognizer.