Avoiding speaker variability in pronunciation verification of children's disordered speech

Saz, Óscar; Lleida, Eduardo; Rodríguez, William Ricardo

doi:10.1145/1640377.1640388

Cited by 3 publications

(1 citation statement)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…More recently, the advancement of machine learning in automatic speech recognition (ASR) has led to a number of Hidden Markov Model (HMM) ASR systems for children's speech evaluations [5,6,7,8,9,10]. The Speech Training, Assessment, and Remediation (STAR) system achieved an r 2 = 0.6 when using phoneme likelihoods in a linear regression to assess the pronunciation of the phoneme /r/ [6].…”

Section: Introductionmentioning

confidence: 99%

Predicting Clinical Evaluations of Children’s Speech with Limited Data Using Exemplar Word Template References

Yeung

Afshan

Ozgun³

et al. 2017

7th ISCA Workshop on Speech and Language Technology in Education (SLaTE 2017)

View full text Add to dashboard Cite

The need for automated speech pathology diagnostic tools for children has increased in recent years. Such tools can help speech pathologists identify speech disorders in children at an early age. This paper introduces an approach to automated clinical evaluations of children's speech using limited data. A database of ten normally developing first-grade children administered the Goldman-Fristoe Test of Articulation, 3rd Edition (GFTA-3) was recorded. Graduate clinicians evaluated the pronunciation of the rhotic sounds by evaluating words in the GFTA-3 containing the letter 'r'. The rhotic sounds were specifically chosen due to their late acquisition in children. Experiments were performed attempting to predict the results of the clinical evaluations. Five children, judged to have proper rhotic pronunciations, were chosen as exemplar templates for the experiment. The remaining children, used for evaluation, were aligned in time to match the five templates using dynamic time warping, and the difference between a test child's 'r' and a template child's 'r' was measured using the cosine distance. Multiple linear regression on the difference scores was shown to be effective at producing predictions that were well-correlated with human clinical evaluations. Several sublists of words with rhotic sounds were used to evaluate the regression, and the sublist containing words with the most mispronunciations performed best. Further discussion includes how much each individual template contributed to the regression and how consistent the clinicians were at scoring children's speech production.

show abstract

Section: Introductionmentioning

confidence: 99%

Predicting Clinical Evaluations of Children’s Speech with Limited Data Using Exemplar Word Template References

Yeung

Afshan

Ozgun³

et al. 2017

7th ISCA Workshop on Speech and Language Technology in Education (SLaTE 2017)

View full text Add to dashboard Cite

show abstract

A Longitudinal Evaluation of Tablet-Based Child Speech Therapy with Apraxia World

Hair

Ballard

Markoulli

et al. 2021

ACM Trans. Access. Comput.

View full text Add to dashboard Cite

Digital games can make speech therapy exercises more enjoyable for children and increase their motivation during therapy. However, many such games developed to date have not been designed for long-term use. To address this issue, we developed Apraxia World, a speech therapy game specifically intended to be played over extended periods. In this study, we examined pronunciation improvements, child engagement over time, and caregiver and automated pronunciation evaluation accuracy while using our game over a multi-month period. Ten children played Apraxia World at home during two counterbalanced 4-week treatment blocks separated by a 2-week break. In one treatment phase, children received pronunciation feedback from caregivers and in the other treatment phase, utterances were evaluated with an automated framework built into the game. We found that children made therapeutically significant speech improvements while using Apraxia World, and that the game successfully increased engagement during speech therapy practice. Additionally, in offline mispronunciation detection tests, our automated pronunciation evaluation framework outperformed a traditional method based on goodness of pronunciation scoring. Our results suggest that this type of speech therapy game is a valid complement to traditional home practice.

show abstract

Assessing Posterior-Based Mispronunciation Detection on Field-Collected Recordings from Child Speech Therapy Sessions

Hair¹,

Zhao²,

Ahmed³

et al. 2021

Interspeech 2021

View full text Add to dashboard Cite

Avoiding speaker variability in pronunciation verification of children's disordered speech

Cited by 3 publications

References 5 publications

Predicting Clinical Evaluations of Children’s Speech with Limited Data Using Exemplar Word Template References

Predicting Clinical Evaluations of Children’s Speech with Limited Data Using Exemplar Word Template References

A Longitudinal Evaluation of Tablet-Based Child Speech Therapy with Apraxia World

Assessing Posterior-Based Mispronunciation Detection on Field-Collected Recordings from Child Speech Therapy Sessions

Contact Info

Product

Resources

About