Acoustic-phonetic transformations for improved speaker-independent isolated word recognition

Zahorian, Stephen A.; Qian, David D.; Jagharghi, Amir J.

doi:10.1109/icassp.1991.150401

Cited by 13 publications

(6 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition, in order to improve the distinguishability and robustness of acoustic features, reduce feature dimensions, and meet the actual needs of continuous speechrecognition systems, researchers have also proposed a variety of feature transformation methods [22][23][24][25][26][27][28][29][30][31][32]; the relevant research is shown in Table 1. LDA, Heteroscedastic Discriminant Analysis (HDA), Generalized Likelihood Ratio Discriminant Analysis (GLRDA), etc., can improve feature discrimination and reduce feature dimensions; Maximum Likelihood Linear Regression (MLLR), fMLLR, and Vocal Tract Length Normalization (VTLN) can eliminate speech information that has nothing to do with the recognition result, such as people or soundtrack, improving the robustness of features.…”

Section: General Approach To Feature Extractionmentioning

confidence: 99%

“…LDA, Heteroscedastic Discriminant Analysis (HDA), Generalized Likelihood Ratio Discriminant Analysis (GLRDA), etc., can improve feature discrimination and reduce feature dimensions; Maximum Likelihood Linear Regression (MLLR), fMLLR, and Vocal Tract Length Normalization (VTLN) can eliminate speech information that has nothing to do with the recognition result, such as people or soundtrack, improving the robustness of features. Discrimination and dimensionality reduction [22] TI/NBS connected digit database [23] A 30-word single syllable highly confusable vocabulary [24] CVC syllables database [25] HAD TI-DIGITS Promote LDA to deal with heteroscedasticity Discrimination and dimensionality reduction [26] GLRDA 200 h of MATBN Mandarin television news…”

Section: General Approach To Feature Extractionmentioning

confidence: 99%

See 1 more Smart Citation

Frontier Research on Low-Resource Speech Recognition Technology

Slam,

Li,

Urouvas

2023

Sensors

View full text Add to dashboard Cite

With the development of continuous speech recognition technology, users have put forward higher requirements in terms of speech recognition accuracy. Low-resource speech recognition, as a typical speech recognition technology under restricted conditions, has become a research hotspot nowadays because of its low recognition rate and great application value. Under the premise of low-resource speech recognition technology, this paper reviews the research status of feature extraction and acoustic models, and conducts research on resource expansion. Especially in terms of the technical challenges faced by this technology, solutions are proposed, and future research directions are prospected.

show abstract

Section: General Approach To Feature Extractionmentioning

confidence: 99%

Section: General Approach To Feature Extractionmentioning

confidence: 99%

Frontier Research on Low-Resource Speech Recognition Technology

Slam,

Li,

Urouvas

2023

Sensors

View full text Add to dashboard Cite

show abstract

“…Linear discriminant analysis is a well-known technique in statistical pattern recognition which has been applied recently to speech recognition to improve recognition performance ͑Brown, 1987; Yu et al, 1990;Zahorian et al, 1991;Hunt and Lefebre, 1989;Umbach and Ney, 1992͒. The idea in linear discriminant analysis ͑LDA͒ is to find a transformation matrix which projects feature vectors from an n-dimensional space to an m-dimensional space (mϽn) such that a class separability criterion is maximized.…”

Section: Introductionmentioning

confidence: 99%

Improving discrimination of confusable words using the divergence measure

Loizou

Spanias

1997

The Journal of the Acoustical Society of America

View full text Add to dashboard Cite

This paper describes the use of the divergence measure as a criterion for finding a transformation matrix which will map the original speech observations onto a subspace with more discriminative ability than the original. A gradient-based algorithm is also proposed to compute the transformation matrix efficiently. The subspace approach is used as a preprocessing step in a hidden Markov model ͑HMM͒ based system to enhance discrimination of acoustically similar pairs of words. This approach is compared with standard linear discriminant analysis ͑LDA͒ techniques and shown to yield as much as 4.5% improvement. The subspace approach is also applied successfully to a more general recognition problem, i.e., discrimination of K confusable words, using the average divergence measure.

show abstract

“…For both training and testing data, the modified Discrete Cosine Transformation Coefficients (DCTC) and Discrete Cosine Series Coefficients (DCSC) (Zahorian et al 1991;Zahorian et al, 1997;Zahorian et al, 2002;Karnjanadecha & Zahorian, 1999) were extracted as original features. The modified DCTC is used for representing speech spectra, and the modified DCSC is used to represent spectral trajectories.…”

Section: Dctc/dcsc Speech Featuresmentioning

confidence: 99%