Automatic speech recognition and a review of its functioning with dysarthric speech

Rosen, Kristin M.; Yampolsky, Sasha

doi:10.1080/07434610012331278904

Cited by 72 publications

(57 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The approach is to convert local Euclidean distances between frame vectors to angles by projecting these -dimensional vectors onto a unit hypersphere centered units from their origin in the st dimension. Namely, every vector is converted to the unit vector sharing an origin with by (7) Given two unit vectors, and that define points on the surface of , the angle between them is by definition (8) Now, given these local distances, we apply symmetric DTW on whole sequences and and get the minimum global distance from the nonlinear aligned Viterbi path with (9) This distance is then converted to the kernel (10) which is symmetric if the symmetric version of DTW is used, which is a requirement for use in SVM classification. In order for the quadratic programming problem to have a definite solution, the kernel must either be a valid dot product [61], or satisfy Mercer's condition, which is to say that given a real-valued kernel , all square integrable functions will give [62].…”

Section: Support Vector Machines (Svms)mentioning

confidence: 99%

See 1 more Smart Citation

Articulatory Knowledge in the Recognition of Dysarthric Speech

Rudzicz

2011

IEEE Trans. Audio Speech Lang. Process.

View full text Add to dashboard Cite

Abstract-Disabled speech is not compatible with modern generative and acoustic-only models of speech recognition (ASR). This work considers the use of theoretical and empirical knowledge of the vocal tract for atypical speech in labeling segmented and unsegmented sequences. These combined models are compared against discriminative models such as neural networks, support vector machines, and conditional random fields. Results show significant improvements in accuracy over the baseline through the use of production knowledge. Furthermore, although the statistics of vocal tract movement do not appear to be transferable between regular and disabled speakers, transforming the space of the former given knowledge of the latter before retraining gives high accuracy. This work may be applied within components of assistive software for speakers with dysarthria.

show abstract

Section: Support Vector Machines (Svms)mentioning

confidence: 99%

“…More commonly, a lack of tongue and lip dexterity often produces heavily slurred speech and a more diffuse and less differentiable vowel target space [7]. The lack of articulatory control often leads to various involuntary sounds caused by velopharyngeal or glottal noise, or noisy swallowing problems [8].…”

Section: A Dysarthriamentioning

confidence: 99%

Articulatory Knowledge in the Recognition of Dysarthric Speech

Rudzicz

2011

IEEE Trans. Audio Speech Lang. Process.

View full text Add to dashboard Cite

show abstract

“…Additionally, phonatory dysfunction and related impairments cause dysarthric speech to be characterized by phonetic distortions, substitutions, and omissions [12,13] that decrease the speaker's intelligibility [1] and thus ASR performance. However it is important to develop ASR systems for dysarthric speakers because of the advantages they offer when compared with interfaces such as switches or keyboards.…”

Section: Introductionmentioning

confidence: 99%

Modelling Errors in Automatic Speech Recognition for Dysarthric Speakers

Morales

Cox

2009

EURASIP J. Adv. Signal Process.

View full text Add to dashboard Cite

Dysarthria is a motor speech disorder characterized by weakness, paralysis, or poor coordination of the muscles responsible for speech. Although automatic speech recognition (ASR) systems have been developed for disordered speech, factors such as low intelligibility and limited phonemic repertoire decrease speech recognition accuracy, making conventional speaker adaptation algorithms perform poorly on dysarthric speakers. In this work, rather than adapting the acoustic models, we model the errors made by the speaker and attempt to correct them. For this task, two techniques have been developed: (1) a set of "metamodels" that incorporate a model of the speaker's phonetic confusion matrix into the ASR process; (2) a cascade of weighted finite-state transducers at the confusion matrix, word, and language levels. Both techniques attempt to correct the errors made at the phonetic level and make use of a language model to find the best estimate of the correct word sequence. Our experiments show that both techniques outperform standard adaptation techniques.

show abstract

“…Instead of relying on another person, speech recognition systems take the burden off of using human transcribers and allow students to independently dictate work (Higgins & Raskind, 1995). Hux, Rankin-Erickson, Manasse, and Lauritzen, (2000) and Rosen and Yampolsky (2000) discussed that due to motor constraints, the hands-free speech recognition system is attractive. It has been suggested that students with physical disabilities could benefit from using speech recognition due to motor difficulties (Duhaney & Duhaney, 2000).…”

Section: Speech Recognition Systemsmentioning

confidence: 99%

“…Because students with physical disabilities often have dysarthric speech, a further investigation explored the use of speech recognition software for this population (Kotler & Thomas-Stonell, 1997). Speech problems that limit the use of speech recognition include decreased intelligibility, phonemic limitations, slower speech rate, variable speech patterns, involuntary sounds, and inconsistency (Kotler & Thomas-Stonell;Rosen & Yampolsky, 2000). Additionally, environmental factors, such as fatigue and time of day need to be considered due to adverse affects on speech (Hux et.…”

Section: Description Of Speech Recognition Software Speech Recognitimentioning

confidence: 99%

Using Speech Recognition Software to Increase Writing Fluency for Individuals with Physical Disabilities

Garrett

Heller

Fowler

et al. 2011

J Spec Educ Technol

View full text Add to dashboard Cite

Students with physical disabilities often have difficulty with writing fluency, despite the use of various strategies, adaptations, and assistive technology (AT). One possible intervention is the use of speech recognition software, although there is little research on its impact on students with physical disabilities. This study used an alternating treatments design to compare the use of speech recognition software to word processing on first draft writing with students with physical disabilities. Areas examined were fluency, accuracy, type of word errors, recall of intended meaning, and length. The results of this study indicated that fluency and length were greater for all participants with speech recognition compared to word processing, but accuracy was lower using speech recognition.

show abstract

Automatic speech recognition and a review of its functioning with dysarthric speech

Cited by 72 publications

References 7 publications

Articulatory Knowledge in the Recognition of Dysarthric Speech

Articulatory Knowledge in the Recognition of Dysarthric Speech

Modelling Errors in Automatic Speech Recognition for Dysarthric Speakers

Using Speech Recognition Software to Increase Writing Fluency for Individuals with Physical Disabilities

Contact Info

Product

Resources

About