Title:
Design of heirarchical perceptron structures and their application to the task of isolated-word recognition
Personal Author:
General Note:
International Joint Conference on Neural Network (USA:June 1989). 243-249
Abstract:
Several design strategies for feed-forward networks are examined within the scope of pattern classification. Single and two-layer perceptron models are adapted for experiments in isolated-word recognition. Direct(one-step) classification as well as several hierarchical(two-step) schemes have been considered. For a vocabulary of twenty English words spoken repeatedly by eleven speakers, the word classes are found to be separable by hyperplanes in the chosen feature space. Since for speaker-dependent word recognition the underlying data base contains only a small training set, an automatic expansion of the training matrials improves the generalization properties of the networks. This method acount for a wide variety of observable improve temporal structures for each word and gives a better overall estimates ofthe network parameters which leads to a recognition rate of 99.5%. For speaker-independent word recognition, a hierarchical structure with pair wise training of two-class models is superior to a single uniform network(98%average recognition rate)
Subject Term:
Added Author:
Available:*
Library | Item Barcode | Call Number | Material Type | Item Category 1 | Status |
---|---|---|---|---|---|
Searching... | 30000000876304 | MAK 3700 | Open Access Book | Article | Searching... |