Title :
Two level cluster analysis based on fractal dimension and iterated function systems (IFS) for speech signal recognition
Author_Institution :
Asian Inst. of Technol., Bangkok, Thailand
Abstract :
A phoneme recognition method is presented based on fractal theory. Fractal dimension and IFS (iterated function system) parameters that represent the methods to calculate the fractal dimension are investigated for word segmentation. The IFS matrices and the eigenvalues of the covariance matrix are proposed for phoneme recognition
Keywords :
covariance matrices; eigenvalues and eigenfunctions; fractals; iterative methods; speech recognition; IFS matrices; covariance matrix; eigenvalues; fractal dimension; iterated function systems; phoneme recognition; speech signal recognition; two level cluster analysis; word segmentation; Acoustic signal detection; Covariance matrix; Eigenvalues and eigenfunctions; Fluctuations; Fractals; Knee; Signal analysis; Size measurement; Speech analysis; Speech recognition;
Conference_Titel :
Circuits and Systems, 1998. IEEE APCCAS 1998. The 1998 IEEE Asia-Pacific Conference on
Conference_Location :
Chiangmai
Print_ISBN :
0-7803-5146-0
DOI :
10.1109/APCCAS.1998.743749
