Consonant characterization using correlation fractal dimension for speech recognition

Langi, Armein Z. R.; Kinsner, W.

doi:10.1109/wescan.1995.493972

Cited by 19 publications

(18 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In any case, the improvement is significant enough so as to be taken into account. In fact, the results confirm the conclusions of previous works, which stated that the most significant benefits of using fractals is their usefulness to differ between voiced and unvoiced sounds [8], and between affricates and other sounds [9]. For example, Figure 1 shows two very similar cases that were mismatched using only MFCC's but where classified correctly using HFD.…”

Section: Results Of the Experimentssupporting

confidence: 88%

“…The interest on fractals in speech date back to the mid-80's [7], and they have been used for a variety of applications, including consonant/vowel characterization [8,9], speaker identification [10], and end-point detection [11], even for whispered speech [12]. Indeed, this metric has been also used in speech recognition, in some cases combined with MFCC's as described above [4].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Combining Mel Frequency Cepstral Coefficients and Fractal Dimensions for Automatic Speech Recognition

Ezeiza

López-de-Ipiña

Hernández

et al. 2011

Advances in Nonlinear Speech Processing

View full text Add to dashboard Cite

Abstract. Hidden Markov Models and Mel Frequency Cepstral Coefficients (MFCC's) are a sort of standard for Automatic Speech Recognition (ASR) systems, but they fail to capture the nonlinear dynamics of speech that are present in the speech waveforms. The extra information provided by the nonlinear features could be especially useful when training data is scarce, or when the ASR task is very complex. In this work, the Fractal Dimension (FD) of the observed time series is combined with the traditional MFCC's in the feature vector in order to enhance the performance of two different ASR systems: the first one is a very simple one, with very few training examples, and the second one is a Large Vocabulary Continuous Speech Recognition System for Broadcast News.

show abstract

Section: Results Of the Experimentssupporting

confidence: 88%

Section: Introductionmentioning

confidence: 99%

Combining Mel Frequency Cepstral Coefficients and Fractal Dimensions for Automatic Speech Recognition

Ezeiza

López-de-Ipiña

Hernández

et al. 2011

Advances in Nonlinear Speech Processing

View full text Add to dashboard Cite

show abstract

“…The interest in fractals in speech date back to the mid-1980s (Pickover and Khorasani, 1986), and they have been used for a variety of applications, including consonant/vowel characterization (Martinez et al, 2003;Langi and Kinsner, 1995), speaker identification (Nelwamondo et al, 2006), and end-point detection (Li et al, 2007), even for whispered speech (Chen and Zhao, 2006). Recent research concerns the analysis of pathological voices through a fractal approach (Chouard et al, 2001;Ouayoun et al, 1999;Péan et al, 2000Péan et al, , 2002.…”

Section: Introductionmentioning

confidence: 99%

Feature selection for spontaneous speech analysis to aid in Alzheimer's disease diagnosis: A fractal dimension approach

López-de-Ipiña

Solé-Casals

Eguiraun

et al. 2015

Computer Speech & Language

View full text Add to dashboard Cite

Alzheimer's disease (AD) is the most prevalent form of degenerative dementia; it has a high socio-economic impact in Western countries. The purpose of our project is to contribute to earlier diagnosis of AD and allow better estimates of its severity by using automatic analysis performed through new biomarkers extracted through non-invasive intelligent methods. The method selected is based on speech biomarkers derived from the analysis of spontaneous speech (SS). Thus the main goal of the present work is feature search in SS, aiming at pre-clinical evaluation whose results can be used to select appropriate tests for AD diagnosis. The feature set employed in our earlier work offered some hopeful conclusions but failed to capture the nonlinear dynamics of speech that are present in the speech waveforms. The extra information provided by the nonlinear features could be especially useful when training data is limited. In this work, the fractal dimension (FD) of the observed time series is combined with linear parameters in the feature vector in order to enhance the performance of the original system while controlling the computational cost. © 2014 Elsevier Ltd. All rights reserved. E-mail addresses: karmele.ipina@ehu.es (K. López-de-Ipiña), jordi.sole@uvic.cat (J. Solé-Casals), harkaitz.eguiraun@ehu.es (H. Eguiraun), jesus.alonso@ulpgc.es (J.B. Alonso), carlos.travieso@ulpgc.es (C.M. Travieso), aitzol.ezeiza@ehu.es (A. Ezeiza), nora.barroso@ehu.es (N. Barroso), mecay@cita-alzheimer.org (M. Ecay-Torres), pmlage@cita-alzheimer.org (P. Martinez-Lage), mariablanca.beitia@ehu.es (B. Beitia).

show abstract

“…However, it can characterize the dynamic properties of the original system for a su ciently large value of n. When the dimension of the underlying state space producing the time series is l, Takens has suggested a value of n¿2l. However, the embedding often works well for smaller values of n [10]. The optimum value of the time delay is still an open question.…”

Section: System Estimation Using State-space Reconstructionmentioning

confidence: 99%

“…Generally, the optimum value for depends on the statistical correlation between the samples. Larger values of can be used when the statistical correlation is high [10].…”

Section: System Estimation Using State-space Reconstructionmentioning

confidence: 99%

Direct use of design criteria in genetic algorithm‐based controller optimization

Kim¹,

Ghaboussi²

2001

Earthq Engng Struct Dyn

View full text Add to dashboard Cite

SUMMARYIn this study, genetic algorithms are applied to design controllers for the wind-excited vibration reduction of a 76-story tall building in the second-generation benchmark problem. Design criteria and constraints including the comfort requirement for building occupants and the control robustness are directly incorporated in the formulation and evaluation of the ÿtness and penalty function of the genetic algorithm-based control design. The method is e cient in obtaining optimal design with multiple constraint conditions. Two controllers are designed with realistic constraints and di erent robustness criteria. The corresponding results are investigated and compared with LQG optimal control results.

show abstract

Consonant characterization using correlation fractal dimension for speech recognition

Cited by 19 publications

References 19 publications

Combining Mel Frequency Cepstral Coefficients and Fractal Dimensions for Automatic Speech Recognition

Combining Mel Frequency Cepstral Coefficients and Fractal Dimensions for Automatic Speech Recognition

Feature selection for spontaneous speech analysis to aid in Alzheimer's disease diagnosis: A fractal dimension approach

Direct use of design criteria in genetic algorithm‐based controller optimization

Contact Info

Product

Resources

About