2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2016
DOI: 10.1109/icassp.2016.7472927
|View full text |Cite
|
Sign up to set email alerts
|
Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
23
0

Year Published

2017
2017
2021
2021

Publication Types

Select...
4
2
1

Relationship

1
6

Authors

Journals

citations
Cited by 37 publications
(23 citation statements)
references
References 13 publications
0
23
0
Order By: Relevance
“…In the first experiment, we evaluate the usefulness of the fine-tuning technique compared to a random initialization of the parameters of the model. The evaluation is based on a measure of intelligibility of the synthesized speech in terms of word accuracy proposed in [19].…”
Section: Methodsmentioning
confidence: 99%
See 1 more Smart Citation
“…In the first experiment, we evaluate the usefulness of the fine-tuning technique compared to a random initialization of the parameters of the model. The evaluation is based on a measure of intelligibility of the synthesized speech in terms of word accuracy proposed in [19].…”
Section: Methodsmentioning
confidence: 99%
“…In this experiment, we synthesized 100 sentences of the Harvard sentences [20] with several models. Then an objective measure of the intelligibility of every sentence was computed in terms of word accuracy [19]. The measure consists of using an ASR to recognize speech and compute a word accuracy by comparing the result to the text label.…”
Section: Objective Measuresmentioning
confidence: 99%
“…The authors developed a model based on Deep Rectifier Neural Networks and Gaussian Processes Regression [18]. In [19] the authors presented a methodology to estimate the neurological state of PD patients from speech signals. Recordings of Spanish, German, and Czech PD patients were considered to estimate their neurological state according to the UPDRS-III score.…”
Section: Assessment Of the Neurological State From Speechmentioning
confidence: 99%
“…Features from Speech-Two different feature sets are considered for speech analysis: The first one is formed with articulation-based features, and includes 86 features such as the energy content in the Bark scale in the transition from voiced to unvoiced segments (22 features), and from unvoiced to voiced segments (22 features) [7]. The feature set is completed with the first and second formant frequencies, and 12 MFCC with their derivatives.…”
Section: Feature Extractionmentioning
confidence: 99%
“…They reported a correlation coefficient of 0.65. In [7], the authors analyzed features related to articulation and intelligibility to predict the UPDRS score of 50 PD patients, and reported a Spearman's correlation of up to 0.72 in a similar scheme as part of the ComParE 2015 challenge. For the gait analysis, in [8], the authors analyzed the influence of gait features to classify PD vs. healthy controls (HC) subjects.…”
Section: Introductionmentioning
confidence: 99%