Are you Anke Meyer-b Ase?

Claim your profile

Publications (1)0 Total impact

  • Source
    Anke Meyer-b Ase · Frank Ohl · Henning Scheich
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper deals with modular radial basis neural networks applying them on stop consonant recognition. The two major areas of concern are the problem of time and the problem of scaling. Speech signals continuously vary over time. To decode these signals we need an appropriate representation of time and we need different kind of networks to fulfill classification tasks. The problem of time is adressed by a recognition system which uses transitional and instantaneous spectral information, the problem of scaling by using different neural classifiers combining fast learning and high performance on recognition tasks and dimensionality reduction methods on the feature space. 1 Introduction
    Preview · Article · Dec 1997