The Automatic Detection of Speech Disorders in Children: Challenges, Opportunities, and Preliminary Results
The DOI matches its Crossref record; 98 of the 102 references it cites were found in a registry or an index.
Abstract
Given the limited accessibility to Speech and Language Pathologists (SLPs) children in need often have, pediatric Computer-Aided Speech Therapy (CAST) tools can play an important role in the early diagnosis and treatment of speech disorders. However, various challenges impede the implementation of accurate automated analysis of speech disorders in children. In this article, we first discuss three key challenges in processing child disordered speech: 1) the unreliability of low-level annotation and scarcity of speech corpora, 2) speaker diarization of therapy sessions and 3) inaccurate children's acoustic models. We next explore opportunities to overcome some of these challenges. First, we investigate the effectiveness of high-level paralinguistic features in disordered speech detection to reduce the dependency on annotated data. A binary classifier trained using paralinguistic features extracted from both typically developing children and those suffering from Speech Sound Disorders (SSD) achieved 87% subject-level classification accuracy. Second, we tackle the speech disorder detection problem as an anomaly detection problem where models are trained merely on typically developing speech, reducing the need for disordered training data. A phoneme-levelF1 score of 0.77 was obtained from an anomaly detection-based system trained on speech attribute features to classify between typical and atypical phoneme pronunciations of children with speech disorder. Finally, we test the efficiency of an x-vector based speaker diarization technique in pediatric therapy sessions. The method successfully distinguished between therapist and child speech with a Diarization Error Rate (DER) of 10%.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“First, and perhaps most critically, there has not been enough publicly available audio or video examples of speech and language disorders to develop AI systems that work well for children or adults with these types of disorders (e.g., Benway et al., 2023;Berisha et al., 2021;Shahin et al., 2020).”
From Equipping Speech-Language Clinicians for the Critical Appraisal of an Artificial Intelligence–Driven, Evidence-Based Future (2025) · cites this paper as Shahin et al., 2020
In the citing paper
“Children’s speech varies significantly across individuals and evolves rapidly, making it difficult for AI models, particularly those trained on adult data, to accurately analyze and detect impairments in younger users. 12”
From Artificial intelligence (AI)-driven technologies for managing pediatric speech and language therapy: A scoping review (2025) · cites this paper as [12]
In the citing paper
“child speech exhibits greater inter- and intra-speaker variability compared to adult speech, posing additional challenges for processing and analysis. 12”
From Artificial intelligence (AI)-driven technologies for managing pediatric speech and language therapy: A scoping review (2025) · cites this paper as [12]
In the citing paper
“As mentioned before, one of the main challenges in working with children’s speech data is the scarcity of available corpora. 12”
From Artificial intelligence (AI)-driven technologies for managing pediatric speech and language therapy: A scoping review (2025) · cites this paper as [12]
In the citing paper
“Another commonly mentioned challenge is the low quality of annotated data for children, primarily due to the wide variability in pronunciations among younger speakers. 12”
From Artificial intelligence (AI)-driven technologies for managing pediatric speech and language therapy: A scoping review (2025) · cites this paper as [12]
Citations
Cites
Gene selection for cancer classification using support vector machines
Guyon, I., Weston, J., Barnhill, S., & Vapnik, V. (2002) · Machine Learning · doi:10.1023/a:1012487302797 · Cited by 9,954 · cited 1×
“nation (RFE) feature selectionmethod [100] to identify the most discriminative parameters.”
BackgroundFront-End Factor Analysis for Speaker Verification
Dehak, N., Kenny, P. J., Dehak, R., Dumouchel, P., & Ouellet, P. (2011) · IEEE Transactions on Audio, Speech, and Language Processing · doi:10.1109/tasl.2010.2064307 · Cited by 3,549
X-Vectors: Robust DNN Embeddings for Speaker Recognition
Snyder, D., Garcia-Romero, D., Sell, G., Povey, D., & Khudanpur, S. (2018) · 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp.2018.8461375 · Cited by 2,620 · cited 3×
“These settings are adopted to match the configurations of the pre-trained model as described in [75].”
MethodSupport Vector Method for Novelty Detection
Schölkopf, B., Williamson, R. C., Smola, A. J., Shawe-Taylor, J. S., & Platt, J. C. (1999) · Cited by 2,040
The Geneva Minimalistic Acoustic Parameter Set (GeMAPS) for Voice Research and Affective Computing
Eyben, F., Scherer, K. R., Schuller, B. W., Sundberg, J., Andre, E., Busso, C., Devillers, L. Y., Epps, J., Laukka, P., Narayanan, S. S., et al. (2016) · IEEE Transactions on Affective Computing · doi:10.1109/taffc.2015.2457417 · Cited by 1,818 · cited 4×
“Here we extracted features in the minimalistic acoustic parameter sets (GeMAPS and eGeMAPS) which include the pitch, jitter, formant frequencies, shimmer, loudness, spectral energies, first four MFCCs, and mean and standard deviation of voiced/unvoiced regions (see [95] for complete list).”
BackgroundMethodInfluentialDARPA TIMIT: acoustic-phonetic continuous speech corpus CD-ROM, NIST speech disc 1-1.1
Garofolo, J. S., Lamel, L. F., Fisher, W. M., Fiscus, J. G., Pallett, D. S., & Dahlgren, N. L. (1993) · doi:10.6028/nist.ir.4930 · Cited by 1,267 · cited 1×
“1) TIMIT Corpus [110]: This is a standard phonetically-rich speech corpus designed for the evaluation of acoustic modeling.”
BackgroundRecent developments in openSMILE, the munich open-source multimedia feature extractor
Eyben, F., Weninger, F., Gross, F., & Schuller, B. (2013) · Proceedings of the 21st ACM international conference on Multimedia · doi:10.1145/2502081.2502224 · Cited by 1,241 · cited 1×
“3 [99] was used for feature extraction.”
The AMI Meeting Corpus: A Pre-announcement
Carletta, J., Ashby, S., Bourban, S., Flynn, M., Guillemot, M., Hain, T., Kadlec, J., Karaiskos, V., Kraaij, W., Kronenthal, M., et al. (2006) · Lecture Notes in Computer Science · doi:10.1007/11677482_3 · Cited by 908 · cited 3×
“While annotating the data, we followed the protocol mentioned in [111].”
BackgroundMethodInfluentialSchuller, B., Steidl, S., Batliner, A., Vinciarelli, A., Scherer, K., Ringeval, F., Chetouani, M., Weninger, F., Eyben, F., Marchi, E., et al. (2013) · Interspeech 2013 · doi:10.21437/interspeech.2013-56 · Cited by 731 · cited 1×
“1) Feature Extraction: The most commonly used standard paralinguistic parameter sets are the large-scale Interspeech 2013 Computational Paralinguistic ChallengE (ComParE) (6,373 parameters) [98], the Geneva Minimalistic Acoustic Parameter Set (GeMAPS) (62 parameters) and its extended version eGeMAPS (88 parameters) [9…”
Speaker Diarization: A Review of Recent Research
Anguera Miro, X., Bozonnet, S., Evans, N., Fredouille, C., Friedland, G., & Vinyals, O. (2012) · IEEE Transactions on Audio, Speech, and Language Processing · doi:10.1109/tasl.2011.2125954 · Cited by 669 · cited 1×
“using factor analysis techniques and shown to performwell [73].”
MethodPhone-level pronunciation scoring and assessment for interactive language learning
Witt, S. M., & Young, S. J. (2000) · Speech Communication · doi:10.1016/s0167-6393(99)00044-8 · Cited by 528
Enhancing one-class support vector machines for unsupervised anomaly detection
Amer, M., Goldstein, M., & Abdennadher, S. (2013) · Proceedings of the ACM SIGKDD Workshop on Outlier Detection and Description · doi:10.1145/2500853.2500857 · Cited by 427 · cited 1×
“The OCSVM operates better when there are no or less anomalies in the training data [107], as in our model where only native speech corpus were used for the training of the anomaly detection model.”
MethodAutomatic Segmentation, Classification and Clustering of Broadcast News Audio
Siegler, M. A. (1997) · Cited by 375
Wang, Q., Downey, C., Wan, L., Mansfield, P. A., & Moreno, I. L. (2018) · 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp.2018.8462628 · Cited by 309
i-vector based speaker recognition on short utterances
Kanagasundaram, A., Vogt, R., Dean, D., Sridharan, S., & Mason, M. (2011) · Interspeech 2011 · doi:10.21437/interspeech.2011-58 · Cited by 234 · cited 1×
“The most common and effective methodologies for diarization involve using i-vectors and DNN embeddings [70]–[72].”
MethodDevelopmental Apraxia of Speech: I. Descriptive and Theoretical Perspectives
Shriberg, L. D., Aram, D. M., & Kwiatkowski, J. (1997) · Journal of Speech, Language, and Hearing Research · doi:10.1044/jslhr.4002.273 · Cited by 220 · cited 1×
“known as developmental verbal dyspraxia, can cause serious communication problems, if not detected and treated in its early stage [25].”
Background(1999) · Cited by 207
Robust recognition of children's speech
Potamianos, A., & Narayanan, S. (2003) · IEEE Transactions on Speech and Audio Processing · doi:10.1109/tsa.2003.818026 · Cited by 203 · cited 3×
“Therefore,acousticmodeladaptationwasusedtoadapt an adult acoustic model to the different age ranges [80].”
BackgroundMethodSpeaker diarization with plda i-vector scoring and unsupervised calibration
Sell, G., & Garcia-Romero, D. (2014) · 2014 IEEE Spoken Language Technology Workshop (SLT) · doi:10.1109/slt.2014.7078610 · Cited by 202 · cited 1×
“[71] used i-vectors along with Probabilistic Linear Discriminant Analysis (PLDA) modelling to score the i-vectors prior to clustering. i-vectors i-vectors are extracted from speaker adapted GMMs using factor analysis techniques and shown to perform well [73].”
MethodArticulation and Phonological Disorders: Speech Sound Disorders in Children
Bernthal, J. E., Bankson, N. W., & Flipsen, P. (2012) · Cited by 192
Shivakumar, P. G., & Georgiou, P. (2020) · Computer Speech & Language · doi:10.1016/j.csl.2020.101077 · Cited by 177
Large-Scale Domain Adaptation via Teacher-Student Learning
Li, J., Seltzer, M. L., Wang, X., Zhao, R., & Gong, Y. (2017) · Interspeech (USB) · doi:10.21437/interspeech.2017-519 · Cited by 160 · cited 1×
“In [91], a teacherstudent domain adaptation approach was employed to adapt a”
MethodThe PF_STAR children's speech corpus
Batliner, A., Blomberg, M., D'Arcy, S., Elenius, D., Giuliani, D., Gerosa, M., Hacker, C., Russell, M., Steidl, S., & Wong, M. (2005) · Interspeech 2005 · doi:10.21437/interspeech.2005-705 · Cited by 158
Tools and Technologies for Computer-Aided Speech and Language Therapy
Saz, Ó., Yin, S.-C., Lleida, E., Rose, R. C., Vaquero, C., & Dueñas, W. R. R. (2009) · Speech Communication · doi:10.1016/j.specom.2009.04.006 · Cited by 153 · cited 3×
“While in Vocaliza [3], the ASR recognized word was compared with the expected one to accept or reject the productions of children with speech dysarthria.”
BackgroundResultStuttering: Clinical and research update.
Perez, H. R., & Stoeckle, J. H. (2016) · Europe PMC (PubMed Central) · Cited by 126
Preston, J. L., Hull, M., & Edwards, M. L. (2013) · American Journal of Speech-Language Pathology · doi:10.1044/1058-0360(2012/12-0022 · Cited by 124
Follow-Up of Children with Early Expressive Phonology Disorders
Lewis, B. A., Freebairn, L. A., & Taylor, H. G. (2000) · Journal of Learning Disabilities · doi:10.1177/002221940003300504 · Cited by 114 · cited 1×
“Delay in their diagnosis and treatment increases the risk of social impairment and learning disabilities [1].”
BackgroundAutomatic speech recognition for children
Potamianos, A., Narayanan, S., & Lee, S. (1997) · 5th European Conference on Speech Communication and Technology (Eurospeech 1997) · doi:10.21437/eurospeech.1997-623 · Cited by 107 · cited 3×
“It has been successfully used to improve performance of child speech recognition by alleviating the acoustic mismatch between child and adult speech [84] and amongst children from different age ranges [85].”
BackgroundMethodClinical Measurement of Stuttering Behaviors
Yaruss, J. S. (1997) · Contemporary Issues in Communication Science and Disorders · doi:10.1044/cicsd_24_s_27 · Cited by 106 · cited 1×
“4) Stuttering: Stuttering is characterized by three main behaviors: 1) blocks, when the child makes a long pause before producing sound, 2) prolongations, when a child stretches out some sounds and 3) repetitions, when a child continuously repeats part of a word or sentence [30], [31] (Fig.”
BackgroundHarrison, A. M., Lo, W.-K., Qian, X.-J., & Meng, H. (2009) · Speech and Language Technology in Education (SLaTE 2009) · doi:10.21437/slate.2009-12 · Cited by 100 · cited 2×
“As these errors are also dominant in second language learners, there is significant research work addressing their automatic detection for Computer Aided Language Learning (CALL) systems [34]–[37].”
BackgroundMethodVocal tract length normalisation approaches to DNN-based children's and adults' speech recognition
Serizel, R., & Giuliani, D. (2014) · IEEE Spoken Language Technology Workshop (SLT) · doi:10.1109/slt.2014.7078563 · Cited by 97 · cited 1×
“Further, VTLN was also effective in improving the phoneme recognition accuracy of both children and adult speech when used with a DNN-HMM acoustic model [86].”
MethodLarge vocabulary automatic speech recognition for children
Liao, H., Pundak, G., Siohan, O., Carroll, M. K., Coccaro, N., Jiang, Q.-M., Sainath, T. N., Senior, A., Beaufays, F., & Bacchiani, M. (2015) · Interspeech 2015 · doi:10.21437/interspeech.2015-373 · Cited by 97 · cited 2×
“Still,themosteffective waytoimprovechildacousticmodelistouselargeamountsof childtrainingdata[81].”
MethodThe OGI kids² speech corpus and recognizers
Shobaki, K., Hosom, J.-P., & Cole, R. A. (2000) · 6th International Conference on Spoken Language Processing (ICSLP 2000) · doi:10.21437/icslp.2000-800 · Cited by 96
Children's speech recognition with application to interactive books and tutors
Hagen, A., Pellom, B., & Cole, R. (2004) · 2003 IEEE Workshop on Automatic Speech Recognition and Understanding (IEEE Cat. No.03EX721) · doi:10.1109/asru.2003.1318426 · Cited by 96 · cited 1×
“It has been successfully used to improve performance of child speech recognition by alleviating the acoustic mismatch between child and adult speech [84] and amongst children from different age ranges [85].”
Speaker Change Detection in Broadcast TV Using Bidirectional Long Short-Term Memory Networks
Yin, R., Bredin, H., & Barras, C. (2017) · Interspeech (USB) · doi:10.21437/interspeech.2017-65 · Cited by 94 · cited 1×
“These normalized coef-ficients were then fed to a deep learning model trained on AMI meeting corpus for speech activity detection [108].”
MethodSiniscalchi, S. M., Lyu, D.-C., Svendsen, T. K., & Lee, C. (2011) · IEEE Transactions on Audio Speech and Language Processing · doi:10.1109/tasl.2011.2167610 · Cited by 85 · cited 1×
“Recognition (LVSR) [104] and universal phoneme recognition [105].”
BackgroundA new DNN-based high quality pronunciation evaluation for computer-aided language learning (CALL)
Hu, W., Qian, Y., & Soong, F. K. (2013) · Interspeech (USB) · doi:10.21437/interspeech.2013-458 · Cited by 82 · cited 2×
“To demonstrate the effectiveness of the method we compared it with the DNN-based GOP algorithm [37].”
MethodAutomatic detection of articulation disorders in children with cleft lip and palate
Maier, A., Hönig, F., Bocklet, T., Nöth, E., Stelzle, F., Nkenke, E., & Schuster, M. (2009) · The Journal of the Acoustical Society of America · doi:10.1121/1.3216913 · Cited by 71 · cited 2×
“The method reported an average detection accuracy of around 89% when compared to manual annotation of Kannada children with CLP. 2) Hypernasality: The automatic detection of hypernasality can be performed on the frame-level [44] or phoneme level [20], [45].”
BackgroundMethodFerrer, L., Bratt, H., Richey, C., Franco, H., Abrash, V., & Precoda, K. (2015) · Speech Communication · doi:10.1016/j.specom.2015.02.002 · Cited by 69 · cited 1×
“[51] proposed a method for lexical stress classification as a part of EduSpeak CALL for English; a GMM model trained using both prosodic and spectral features was employed to classify each syllable as unstressed, primary or secondary.”
MethodImproving Children’s Speech Recognition Through Out-of-Domain Data Augmentation
Fainberg, J., Bell, P., Lincoln, M., & Renals, S. J. (2016) · Interspeech (USB) · doi:10.21437/interspeech.2016-1348 · Cited by 67 · cited 1×
“In [87], Stochastic Feature Mapping (SFM) was employed to augment child speech by randomly transforming features from one of the out-of-domain (adult) speakers to one of the in-domain (child) speakers using feature space Maximum Like-lihood Linear Regression (fMLLR).”
MethodSpeech-driven mobile games for speech therapy: User experiences and feasibility
Ahmed, B., Monroe, P., Hair, A., Tan, C. T., Gutierrez-Osuna, R., & Ballard, K. J. (2018) · International Journal of Speech-Language Pathology · doi:10.1080/17549507.2018.1513562 · Cited by 65 · cited 1×
“Only STAR, Vocaliza and Tabby Talks utilize speech analysis to provide automated feedback and help guide the therapy exercises [11].”
BackgroundSelf-citationOn the Difficulties of Automatic Speech Recognition for Kindergarten-Aged Children
Yeung, G., & Alwan, A. (2018) · Interspeech 2018 · doi:10.21437/interspeech.2018-2297 · Cited by 64 · cited 1×
“[89] were used.”
UltraSuite: A Repository of Ultrasound and Acoustic Data from Child Speech Therapy Sessions
Eshky, A., Ribeiro, M. S., Cleland, J., Richmond, K., Roxburgh, Z., Scobbie, J. M., & Wrench, A. (2018) · Interspeech 2018 · doi:10.21437/interspeech.2018-1736 · Cited by 64 · cited 3×
“4) UltraSuite Corpus [24]: This is a speech recordings and ultrasound repository collected from 58 TD children (ages: 9y 3m ± 1y 10m) and 28 children (ages: 8y 2m ± 2y) with different typesofSSDsincluding,childhoodapraxiaofspeech,phonolog-icaldelayandphonologicalandarticulationdisorders.”
BackgroundMethodInfluentialComputer aided pronunciation learning system using speech recognition techniques
Abdou, S. M., Hamid, S. E., Rashwan, M., Samir, A., Abdel‐Hamid, O., Shahin, M. A., & Nazih, W. (2006) · Interspeech (USB) · doi:10.21437/interspeech.2006-287 · Cited by 60
Self-citationPitch-Adaptive Front-End Features for Robust Children’s ASR
Shahnawazuddin, S., Dey, A., & Sinha, R. (2016) · Interspeech (USB) · doi:10.21437/interspeech.2016-1020 · Cited by 60 · cited 2×
“PAMFCCs were proposed in [50] to reduce the effect of the increase of pitch and formant frequencies on the extracted features and produce pitch-robust MFCC features.”
BackgroundSTAR: articulation training for young children
Bunnell, H. T., Yarrington, D. M., & Polikoff, J. B. (2000) · 6th International Conference on Spoken Language Processing (ICSLP 2000) · doi:10.21437/icslp.2000-757 · Cited by 58 · cited 2×
“These include Speech Training, Assessment, and Remediation (STAR) [2], Vocaliza [3], Speech Assessment and Interactive Learning System (SAILS) [4], Phoneme Factory Sound Sorter (PFSS) [5], Tabby Talks [6], [7], Apraxiaville [8], ArtikPix [9] and Apraxia World [10].”
BackgroundMethodLee, C., & Siniscalchi, S. M. (2013) · Proceedings of the IEEE · doi:10.1109/jproc.2013.2238591 · Cited by 56 · cited 2×
“compared to traditional features [102].”
BackgroundTabby Talks: An automated tool for the assessment of childhood apraxia of speech
Shahin, M. A., Ahmed, B., Parnandi, A., Karappa, V., McKechnie, J., Ballard, K. J., & Gutiérrez‐Osuna, R. (2015) · Speech Communication · doi:10.1016/j.specom.2015.04.002 · Cited by 51
Self-citationUniversal attribute characterization of spoken languages for automatic spoken language recognition
Siniscalchi, S. M., Reed, J. T., Svendsen, T. K., & Lee, C. (2012) · Computer Speech & Language · doi:10.1016/j.csl.2012.05.001 · Cited by 49 · cited 1×
“Moreover, they have been successfully utilized in a various of speech processing domains including identification of spoken language [103], lattice rescoring for Large Vocabulary Speech Recognition (LVSR) [104] and universal phoneme recognition [105].”
Golabbakhsh, M., Abnavi, F., Kadkhodaei Elyaderani, M., Derakhshandeh, F., Khanlar, F., Rong, P., & Kuehn, D. P. (2017) · The Journal of the Acoustical Society of America · doi:10.1121/1.4976056 · Cited by 49
Apraxia world: a speech therapy game for children with speech sound disorders
Hair, A., Monroe, P., Ahmed, B., Ballard, K. J., & Gutierrez-Osuna, R. (2018) · Proceedings of the 17th ACM Conference on Interaction Design and Children · doi:10.1145/3202185.3202733 · Cited by 45 · cited 1×
“These include Speech Training, Assessment, and Remediation (STAR) [2], Vocaliza [3], Speech Assessment and Interactive Learning System (SAILS) [4], Phoneme Factory Sound Sorter (PFSS) [5], Tabby Talks [6], [7], Apraxiaville [8], ArtikPix [9] and Apraxia World [10].”
BackgroundSelf-citationIntervention for dysarthria associated with acquired brain injury in children and adolescents
Morgan, A. T., & Vogel, A. P. (2008) · Cochrane Database of Systematic Reviews · doi:10.1002/14651858.cd006279.pub2 · Cited by 44 · cited 1×
“Mild hypernasality and/or hoarse voice are also common in dysarthric speech [28].”
BackgroundPerceptual and acoustic reliability estimates for the Speech Disorders Classification System (SDCS)
Shriberg, L. D., Fourakis, M. S., Hall, S. D., Karlsson, H. B., Lohmeier, H. L., McSweeny, J. L., Potter, N. L., Scheer-Cohen, A. R., Strand, E. A., Tilkens, C. M., et al. (2010) · Clinical Linguistics & Phonetics · doi:10.3109/02699206.2010.503007 · Cited by 44
Automatic Classification of Autistic Child Vocalisations: A Novel Database and Results
Baird, A., Amiriparian, S., Cummins, N., Alcorn, A. M., Batliner, A. M., Pugachevskiy, S., Freitag, M., Gerczuk, M., & Schuller, B. W. (2017) · Interspeech (USB) · doi:10.21437/interspeech.2017-730 · Cited by 43 · cited 1×
“Paralinguistic acoustic parameters are low-level descriptors of the prosodic, spectral and voice quality of speech, typically used to analyze emotion in speech [95] and detect autism spectrum disorder (ASD) [96], [97].”
BackgroundOrozco-Arroyave, J. R., Murillo-Rendón, S., Álvarez-Meza, A. M., Arias-Londoño, J. D., Delgado-Trejos, E., Vargas-Bonilla, J. F., & Castellanos-Domínguez, C. G. (2011) · Interspeech 2011 · doi:10.21437/interspeech.2011-223 · Cited by 42 · cited 1×
“Most hypernasality detectors perform binary classification to determine the existence or absence of hypernasality [20], [45]–[47].”
BackgroundAn exploration of the role of executive functions in preschoolers’ phonological development
Eaton, C. T., & Ratner, N. B. (2016) · Clinical Linguistics & Phonetics · doi:10.1080/02699206.2016.1179344 · Cited by 41
Automatic analysis of pronunciations for children with speech sound disorders
Dudy, S., Bedrick, S., Asgari, M., & Kain, A. (2018) · Computer Speech & Language · doi:10.1016/j.csl.2017.12.006 · Cited by 38 · cited 1×
“Dudy et al. [17] incorporated an ERN in the computation of the Goodness of Pronunciation (GOP) derived from the mispronunciations of children with articulation disorders and autism.”
MethodA comparison of word lexicality in the treatment of speech sound disorders
Cummings, A. E., & Barlow, J. A. (2010) · Clinical Linguistics & Phonetics · doi:10.3109/02699206.2010.528822 · Cited by 36
Development of a Remote Therapy Tool for Childhood Apraxia of Speech
Parnandi, A., Karappa, V., Lan, T., Shahin, M., McKechnie, J., Ballard, K., Ahmed, B., & Gutierrez-Osuna, R. (2015) · ACM Transactions on Accessible Computing · doi:10.1145/2776895 · Cited by 34 · cited 1×
“These include Speech Training, Assessment, and Remediation (STAR) [2], Vocaliza [3], Speech Assessment and Interactive Learning System (SAILS) [4], Phoneme Factory Sound Sorter (PFSS) [5], Tabby Talks [6], [7], Apraxiaville [8], ArtikPix [9] and Apraxia World [10].”
BackgroundSelf-citationShahin, M., Ahmed, B., McKechnie, J., Ballard, K., & Gutierrez-Osuna, R. (2014) · Interspeech 2014 · doi:10.21437/interspeech.2014-377 · Cited by 34 · cited 1×
“Deep Neural Network (DNN-HMM)-based [39].”
Self-citationAutomatic Detection of Hypernasality in Children
Rendón, S. M., Orozco‐Arroyave, J. R., Vargas‐Bonilla, J. F., Arias-Londoño, J. D., & Castellanos-Dominguez, G. (2011) · Lecture notes in computer science · doi:10.1007/978-3-642-21326-7_19 · Cited by 33 · cited 3×
“In[45],measuresof voice quality including jitter, shimmer and harmonic-to-noise ratio were utilized along with MFCCs to perform a binary classification of Spanish children vowel production as normal or hypernasal.”
BackgroundMethodDeep neural network adaptation for children's and adults' speech recognition
Serizel, R., & Giuliani, D. (2014) · Pisa University Press eBooks · doi:10.12871/clicit2014166 · Cited by 33 · cited 2×
“In [92], transfer learning from adult to child DNN”
BackgroundA Lightly Supervised Approach to Detect Stuttering in Children's Speech
Alharbi, S., Hasan, M., Simons, A. J. H., Brumfitt, S., & Green, P. (2018) · Interspeech 2018 · doi:10.21437/interspeech.2018-2155 · Cited by 33
Lexical stress detection for L2 English speech using deep belief networks
Li, K., Qian, X., Kang, S., & Meng, H. M. L. (2013) · Interspeech (USB) · doi:10.21437/interspeech.2013-447 · Cited by 30 · cited 1×
“HMM [54], Maximum Entropy (MaxEnt) [55] and Deep Belief Network (DBN) [56] have also been employed to perform lexical stress classification.”
MethodAutomatic Evaluation of Hypernasality Based on a Cleft Palate Speech Database
He, L. T., Zhang, J., Liu, Q., Yin, H., Lech, M., & Huang, Y. (2015) · Journal of Medical Systems · doi:10.1007/s10916-015-0242-2 · Cited by 28 · cited 1×
“In [48] a four-way k-Nearest Neighbor (kNN) classifier was proposed to classify hypernasal speech as normal, moderate, mild or severe.”
Improving Children Speech Recognition through Feature Learning from Raw Speech Signal
Dubagunta, S. P., Kabil, S. H., & Magimai.-Doss, M. (2019) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2019.8682826 · Cited by 27 · cited 1×
“Most recently, in [88] the authors examined the effectiveness of applying deep learning to learn useful features directly from the raw speech signal instead of using hand-crafted features such”
BackgroundExploiting Acoustic and Syntactic Features for Prosody Labeling in a Maximum Entropy Framework
Sridhar, V. K. R., Bangalore, S., & Narayanan, S. S. (2007) · North American Chapter of the Association for Computational Linguistics · Cited by 26
Bernhardt, B. M., Hanson, R., Pérez, D., Ávila, C., Lleó, C., Stemberger, J. P., Carballo, G., Mendoza, E., Fresneda, D., & Chávez-Peón, M. E. (2014) · International Journal of Language & Communication Disorders · doi:10.1111/1460-6984.12133 · Cited by 26
Automatic lexical stress detection using acoustic features for computer-assisted language learning
Zhao, J., Yuan, H., Liu, J., & Xia, S. (2021) · Open Access Victoria University of Wellington | Te Herenga Waka (Figshare) · doi:10.26686/wgtn.14524884.v1 · Cited by 25
Estimation of Hypernasality Scores from Cleft Lip and Palate Speech
Vikram, C. M., Tripathi, A., Kalita, S., & Prasanna, S. R. M. (2018) · Interspeech (USB) · doi:10.21437/interspeech.2018-1631 · Cited by 25 · cited 1×
“[19]–[23] and dysarthria.”
The ACLEW DiViMe: An Easy-to-use Diarization Tool
Franc, A. L., Riebling, E., Karadayi, J., Wang, Y., Scaff, C., Metze, F., & Cristià, A. (2018) · Interspeech (USB) · doi:10.21437/interspeech.2018-2324 · Cited by 24 · cited 3×
“Withinthedomainofchildren’sspeech,thelimitedworkdone todatehasshownthatitisanon-trivialtask[76].”
BackgroundMethodPronunciation analysis for children with speech sound disorders
Dudy, S., Asgari, M., & Kain, A. (2015) · 2015 37th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc.2015.7319655 · Cited by 22 · cited 2×
“There are four main pronunciation errors associated with articulation and phonological disorders [15]: 1) phoneme replacement, i.e. replacing certain phoneme with another, 2) phoneme distortion, i.e. altering acoustic characteristics of one or more phonemes, 3) phoneme deletion and 4) phoneme…”
BackgroundAutomatic lexical stress detection for Chinese learners' of English
Chen, J., & Wang, L. (2010) · International Symposium on Chinese Spoken Language Processing (ISCSLP) · doi:10.1109/iscslp.2010.5684859 · Cited by 21 · cited 1×
“[52] utilized differential features computed from the vowels of consecutive syllables in a SVM classifier to detect the lexical stress pattern in the speech productions of Chinese learners of English.”
MethodEnglish sentence stress detection system based on HMM framework
Li, C., Liu, J., & Xia, S. (2006) · Applied Mathematics and Computation · doi:10.1016/j.amc.2006.06.081 · Cited by 20 · cited 7×
“HMM [54], Maximum Entropy (MaxEnt) [55] and Deep Belief Network (DBN) [56] have also been employed to perform lexical stress classification.”
MethodInfluentialTowards Cross-lingual Automatic Diagnosis of Autism Spectrum Condition in Children's Voices.
Schmitt, M., Marchi, E., Ringeval, F., & Schuller, B. W. (2016) · OPUS (Augsburg University) · Cited by 20
Automatic Identification of Articulation Disorders for Arabic Children Speakers
Hanani, A. A., Attari, M., Farakhna, A., Joma'A, A., Hussein, M. S., & Taylor, S. (2016) · Workshop on Child, Computer and Interaction (WOCCI) · doi:10.21437/wocci.2016-6 · Cited by 19 · cited 3×
“A similar approach was adopted in [43] to detect mispronunciations of a single Arabic phoneme at”
BackgroundMethodSpeaker independent diarization for child language environment analysis using deep neural networks
Najafian, M., & Hansen, J. H. L. (2016) · IEEE Spoken Language Technology Workshop (SLT) · doi:10.1109/slt.2016.7846253 · Cited by 19 · cited 1×
“[79] proposed a HMM based speaker diarization system with DNNs replacing GMMs for probability estimation and reported results on private data collected at a children day-care-center.”
Improving Child Speech Disorder Assessment by Incorporating Out-of-Domain Adult Speech
Smith, D., Sneddon, A., Ward, L., Duenser, A., Freyne, J., Silvera‐Tawil, D., & Morgan, A. T. (2017) · Interspeech (USB) · doi:10.21437/interspeech.2017-455 · Cited by 17 · cited 3×
“…of child disordered speech, to deal with the shortage of speech corpora and unreliability of low-level annotation including incorporating out-of-domain adult speech [16], extracting features on speaker-level such as i-vectors [41]–[43] and using robust acoustic features such as pause events [69].”
BackgroundMethodDetecting Stuttering Events in Transcripts of Children’s Speech
Alharbi, S., Hasan, M., Simons, A. J. H., Brumfitt, S. M., & Green, P. (2017) · Lecture notes in computer science · doi:10.1007/978-3-319-68456-7_18 · Cited by 17 · cited 1×
“There is some recent research on the automatic detection of the different types of children speech disorders including stuttering [12]–[14], speech sound disorders [15]–[18], hypernasality”
BackgroundPitch-Adaptive Front-end Feature for Hypernasality Detection
Dubey, A. K., Prasanna, S. R. M., & Dandapat, S. (2018) · Interspeech (USB) · doi:10.21437/interspeech.2018-1251 · Cited by 16 · cited 4×
“2) Hypernasality: The automatic detection of hypernasality can be performed on the frame-level [44] or phoneme level [20], [45].”
BackgroundMethodInfluentialWard, L., Stefani, A., Smith, D., Duenser, A., Freyne, J., Dodd, B., & Morgan, A. (2016) · Interspeech 2016 · doi:10.21437/interspeech.2016-850 · Cited by 14 · cited 1×
“Ward et al. [40] constructed a lattice for the most common error patterns of children with different types of SSDs and decoded it with Hierarchical Neural Network (HNN) acoustic model.”
MethodA Study on Acoustic Modeling for Child Speech Based on Multi-Task Learning
Wang, J., Ng, S. I., Tao, D., Ng, W. Y., & Lee, T. (2018) · International Symposium on Chinese Spoken Language Processing (ISCSLP) · doi:10.1109/iscslp.2018.8706703 · Cited by 14 · cited 1×
“A multi-task learning approach with child speech as a primary task and adult speech as a secondary task was adopted in [94] to train CNN and TDNN acoustic models.”
Methodde Antioquia, U., Orozco‐Arroyave, J. R., Vargas‐Bonilla, J. F., de Antioquia, U., Vásquez-Correa, J. C., de Antioquia, U., Castellanos-Domínguez, C. G., Manizales), U. N. D. C. (., & Nöth, E. (2016) · Revista Facultad de Ingeniería Universidad de Antioquia · doi:10.17533/udea.redin.n80a12 · Cited by 13 · cited 3×
“It is important to note that [19]”
BackgroundClassification of lexical stress patterns using deep neural network architecture
Shahin, M. A., Ahmed, B., & Ballard, K. J. (2014) · IEEE Spoken Language Technology Workshop (SLT) · doi:10.1109/slt.2014.7078621 · Cited by 12 · cited 1×
“In the context of child disordered speech, we trained a DNN model using both differential and raw acoustic features to detect lexical stress errors in children with CAS [58].”
MethodSelf-citationKothalkar, P. V., Rudolph, J. M., Dollaghan, C. A., McGlothlin, J., Campbell, T. F., & Hansen, J. H. L. (2018) · Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc.2018.8513447 · Cited by 11 · cited 1×
“There is some recent research on the automatic detection of the different types of children speech disorders including stuttering [12]–[14], speech sound disorders [15]–[18], hypernasality”
BackgroundAttribute based lattice rescoring in spontaneous speech recognition
Chen, I.-F., Siniscalchi, S. M., & Lee, C. (2014) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2014.6854216 · Cited by 10 · cited 1×
“Moreover, they have been successfully utilized in a various of speech processing domains including identification of spoken language [103], lattice rescoring for Large Vocabulary Speech Recognition (LVSR) [104] and universal phoneme recognition [105].”
Wren, Y. E., Roulstone, S., & Williams, A. L. (2010) · Digital Commons - East Tennessee State University (East Tennessee State University) · Cited by 10
Guo, J., Paturi, R., Yeung, G., Lulich, S. M., Arsikere, H., & Alwan, A. A. (2015) · Interspeech (USB) · doi:10.21437/interspeech.2015-384 · Cited by 9 · cited 1×
“in [90] using subglottal resonances.”
Towards an Automated Screening Tool for Developmental Speech and Language Impairments
Gong, J. J., Gong, M., Levy-Lambert, D., Green, J. R., Hogan, T. P., & Guttag, J. V. (2016) · Interspeech (USB) · doi:10.21437/interspeech.2016-549 · Cited by 8 · cited 1×
“of child disordered speech, to deal with the shortage of speech corpora and unreliability of low-level annotation including incorporating out-of-domain adult speech [16], extracting features on speaker-level such as i-vectors [41]–[43] and using robust acoustic features such as pause events [69].”
BackgroundVikram, C. M., Prasanna, S. R. M., Abraham, A. K., Pushpavathi, M., & Girish, K. S. (2018) · Interspeech (USB) · doi:10.21437/interspeech.2018-1665 · Cited by 8 · cited 2×
“In a method to detect GAE proposed by Vikram et al [22], the speech was first passed through a glottal activity detector to detect regions where the glottal vibration exists, followed by two detectors for syllable vowels and voiced consonants respectively.”
BackgroundMethodFusing Text-dependent Word-level i-Vector Models to Screen ‘at Risk’ Child Speech
Kothalkar, P. V., Rudolph, J. M., Dollaghan, C. A., McGlothlin, J., Campbell, T. F., & Hansen, J. H. L. (2018) · Interspeech (USB) · doi:10.21437/interspeech.2018-1465 · Cited by 8 · cited 3×
“A speaker verification-based method using i-vector features was also applied to discriminate between typically developing (TD) and SSD speakers [41]–[43].”
BackgroundMethodA Progressive Deep Learning Approach to Child Speech Separation
Wang, X., Du, J., Sun, L., Wang, Q., & Lee, C. (2018) · International Symposium on Chinese Spoken Language Processing (ISCSLP) · doi:10.1109/iscslp.2018.8706689 · Cited by 6 · cited 2×
“[78] presented a stacked RNN architecture to separate child and adult speech by training the neural network on multiple objective functions and then jointly training the shared neural network with results.”
MethodBergelson, E., Warlaumont, A. S., Cristià, A., Casillas, M., Rosemberg, C. R., Soderstrom, M., Rowland, C., Durrant, S., & Bunce, J. (2017) · Databrary · doi:10.17910/b7.390 · Cited by 6
Automatic detection of articulations disorders from children’s speech preliminary study
Ramou, N., & Guerti, M. (2014) · Journal of Communications Technology and Electronics · doi:10.1134/s1064226914110187 · Cited by 5
Classification of bisyllabic lexical stress patterns in disordered speech using deep learning
Shahin, M. A., Gutiérrez‐Osuna, R., & Ahmed, B. (2016) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2016.7472925 · Cited by 5 · cited 2×
“In our previous work, we proposed a Convolutional Neural Network (CNN) based classifier to discriminate between primary stressed, secondary stressed and unstressed syllables and applied it to English and Arabic speech [57].”
BackgroundMethodSelf-citationShahin, M. A., Ahmed, B., Ji, J., & Ballard, K. J. (2018) · Interspeech (USB) · doi:10.21437/interspeech.2018-1319 · Cited by 5 · cited 1×
“the automatic detection of disordered speech as an anomaly detection problem [101].”
BackgroundSelf-citationSelf-similarity Matrix Based Intelligibility Assessment of Cleft Lip and Palate Speech
Kalita, S., Prasanna, S. R. M., & Dandapat, S. (2018) · Interspeech (USB) · doi:10.21437/interspeech.2018-1125 · Cited by 4
Disfluent Speech Segments Detection and Remediation.
Arbajian, P. (2019) · ProQuest LLC eBooks · Cited by 1
Research works printed but not linked: 7
- chee, 0, Overview of automatic stuttering recognition system, Proc Int Conf Man-Mach Syst
- al-barhamtoshy, 2014, Speak correct: Phonetic editor approach, Life Sci, 11, 626
- vikramc, 0, Estimation of hypernasality scores from cleft lip and palate speech, Proc INTERSPEECH, 1701
- 2007, Childhood Apraxia of Speech Position Statement
- rvachew, 2010, Speech perception intervention, Interventions for Speech Sound Disorders in Children, 295
- chiat, 1994, From lexical access to lexical output: What is the problem for children with impaired phonology, First Second Lang Pathol, 107
- eskenazi, 1997, The CMU Kids Speech Corpus
Cited by
Informing Age-Appropriate AI: Examining Principles and Practices of AI for Children
Wang, G., Zhao, J., Van Kleek, M., & Shadbolt, N. (2022) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3491102.3502057 · Cited by 76
Dysarthric-speech detection using transfer learning with convolutional neural networks
Sekhar, S. R. M., Kashyap, G., Bhansali, A., Azhagesan, A. A. A., & Singh, K. (2021) · ICT Express · doi:10.1016/j.icte.2021.07.004 · Cited by 53
Beccaluva, E. A., Catania, F., Arosio, F., & Garzotto, F. (2024) · Human–Computer Interaction · doi:10.1080/07370024.2023.2242837 · Cited by 24
“…set instead of a larger or different set as eGeMAPS offers some advantages: it reduces the possibility of overfitting our models to the training data; moreover, the feature set has already been shown as promising for assessing linguistic disorders (Shahin et al., 2019, 2020; Xue et al., 2019).”
BackgroundJothi, K. R., Sivaraju, S. S., & Yawalkar, P. J. (2020) · International Conference on Electronics, Communication and Aerospace Technology (ICECA) · doi:10.1109/iceca49313.2020.9297591 · Cited by 23
“2020), focused on outstanding approach used to find the 3 common pronunciation errors made by children [31].”
BackgroundMethodDeveloping children's ASR system under low-resource conditions using end-to-end architecture
Ankita, & Shahnawazuddin, S. (2024) · Digital Signal Processing · doi:10.1016/j.dsp.2024.104385 · Cited by 21
Metadata
- Authors
- Shahin, M., Zafar, U., & Ahmed, B.
- Affiliations
- UNSW Sydney (AU)
- Texas A&M University at Qatar (QA)
- Venue
- IEEE Journal of Selected Topics in Signal Processing
- Published
- 2020-02
- Volume
- 14
- Issue
- 2
- Pages
- 400-412
- ISSN
- 1932-4553 (ISSN-L)
- 1941-0484
- OpenAlex
- W3016223211
- Cited by
- 62
- References
- 102
- Citations by year
- 2026: 9 · 2025: 14 · 2024: 14 · 2023: 7 · 2022: 10 · 2021: 3 · 2020: 5
- FWCI
- 2.08
- Type
- journal article
- Language
- en
- Open access
- closed
- Licence
- Locations
- IEEE Journal of Selected Topics in Signal Processing (publishedVersion)
- Funders
- NPRP (8-293-2-124)
- Qatar National Research Fund
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Speech Recognition and Synthesis
- Social Sciences › Psychology › Developmental and Educational Psychology › Language Development and Disorders
- Health Sciences › Health Professions › Speech and Hearing › Voice and Speech Disorders
- Keywords
- children, speech sound disorders, paralinguistic features, anomaly detection, acoustic modeling, phoneme classification
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.