1983
DOI: 10.1002/j.1538-7305.1983.tb03115.x
|View full text |Cite
|
Sign up to set email alerts
|

On the Application of Vector Quantization and Hidden Markov Models to Speaker-Independent, Isolated Word Recognition

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
70
0
2

Year Published

1990
1990
2023
2023

Publication Types

Select...
4
3

Relationship

0
7

Authors

Journals

citations
Cited by 273 publications
(72 citation statements)
references
References 11 publications
0
70
0
2
Order By: Relevance
“…In our experiments, a constrained left-to-right HMM structure with five states, as described by Rabiner [3], was used. The model is shown in Fig.…”
Section: Experiments and Resultsmentioning
confidence: 99%
See 2 more Smart Citations
“…In our experiments, a constrained left-to-right HMM structure with five states, as described by Rabiner [3], was used. The model is shown in Fig.…”
Section: Experiments and Resultsmentioning
confidence: 99%
“…In the LBG case, larger codebooks require greater amount of computation, but yield comparatively small gains in recognition accuracy. Rabiner [3] suggests a codebook size of 64 for the isolated digits recognition task. In our experience a codebook with 128 members has been found to represent a reasonable balance between the amount of computation required and the resulting recognition accuracy.…”
Section: Experiments and Resultsmentioning
confidence: 99%
See 1 more Smart Citation
“…u nt log u nt (13) where d nt is redefined as in (8), and m E > 0 is a value which controls the degree of fuzzy entropy. Minimizing the H-function with respect to λ results in the following FE-GMM algorithm 1) Given a training data set X = {x 1 , x 2 , .…”
Section: Minimum Fuzzy Entropy Squared-error Estimationmentioning
confidence: 99%
“…It was shown in the literature that no significant difference in pattern recognition was found by using different initialisation methods [13], [14]. Therefore the following parameters of the GMM are initialised: Mixture weights, mean vectors, covariance matrices and fuzzy membership functions.…”
Section: Initialisation and Constraints On Parameters During Trainingmentioning
confidence: 99%