Abstract
Phoneme segmentation is important for many healthcare applications, such as the diagnosis and monitoring of children with speech sound disorders (SSDs). This is usually addressed by performing forced alignment (FA), which essentially annotates an audio file to provide information on what has been uttered and where. While many FA tools exist, very few can work automatically without the assistance of a transcription. This work aims at providing a novel text-independent FA tool by using two models, namely wav2vec 2.0 and an unsupervised segmentor known as UnsupSeg. To provide labels to the segments, the class regions that are obtained by nearest-neighbour classification with wav2vec 2.0 labels pre-CTC collapse as the reference points. Maximal overlap between the class regions and the segments determines class label. Additional post-processing steps, such as over-fitting cleaning and application of voice activity detection, are also performed to further improve the segmentation performance. All the models used to create the tool are self-supervised, and thus can leverage great amounts of unlabelled data to reduce the need for labelled data. When evaluated on the TIMIT dataset, our implementation achieved a harmonic mean score of 76.88%, competitive against other alternatives.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
This refers to conditional dependence on the previous states in a sequence.
References
Speech sound disorders-articulation and phonology (2022). https://www.asha.org/practice-portal/clinical-topics/articulation-and-phonology/
Baevski, A., Zhou, Y., Mohamed, A., Auli, M.: wav2vec 2.0: a framework for self-supervised learning of speech representations. In: Proceedings of Advances in Neural Information Processing Systems, vol. 33, pp. 12449–12460 (2020)
Daniel, G.R., McLeod, S.: Children with speech sound disorders at school: challenges for children, parents and teachers. Aust. J. Teach. Educ. 42(2), 81–101 (2017)
Furlong, L., Serry, T., Erickson, S., Morris, M.E.: Processes and challenges in clinical decision-making for children with speech-sound disorders. Int. J. Lang. Commun. Disord. 53(6), 1124–1138 (2018)
Garofolo, J.S., et al.: TIMIT Acoustic-Phonetic Continuous Speech Corpus (1993). https://catalog.ldc.upenn.edu/LDC93S1
Graves, A., Schmidhuber, J.: Framewise phoneme classification with bidirectional LSTM and other neural network architectures. Neural Netw. 18(5–6), 602–610 (2005)
Konstantinidis, S.: Computing the edit distance of a regular language. Inf. Comput. 205(9), 1307–1316 (2007)
Kreuk, F.: Self-Supervised Contrastive Learning for Unsupervised Phoneme Segmentation (INTERSPEECH 2020) (2021)
Kreuk, F., Keshet, J., Adi, Y.: Self-Supervised Contrastive Learning for Unsupervised Phoneme Segmentation. arXiv:2007.13465 [cs, eess, stat] (2020)
Kreuk, F., Sheena, Y., Keshet, J., Adi, Y.: Phoneme boundary detection using learnable segmental features. In: Proceedings of ICASSP, Barcelona, Spain, pp. 8089–8093. IEEE (2020)
Mahr, T., Berisha, V., Kawabata, K., Liss, J., Hustad, K.: Performance of forced-alignment algorithms on children’s speech. Technical report, PsyArXiv (2020)
McKechnie, J., Ahmed, B., Gutierrez-Osuna, R., Monroe, P., McCabe, P., Ballard, K.J.: Automated speech analysis tools for children’s speech production: a systematic literature review. Int. J. Speech Lang. Pathol. 20(6), 583–598 (2018)
McKechnie, J.G.: Exploring the use of technology for assessment and intensive treatment of childhood apraxia of speech. Ph.D. thesis (2019)
Mcleod, S., Baker, E.: Speech-language pathologists’ practices regarding assessment, analysis, target selection, intervention, and service delivery for children with speech sound disorders. Clin. Linguist. Phon. 28(7–8), 508–531 (2014)
McLeod, S., et al.: Profile of Australian preschool children with speech sound disorders at risk for literacy difficulties. Aust. J. Learn. Diffic. 22(1), 15–33 (2017)
McLeod, S., et al.: Tutorial: speech assessment for multilingual children who do not speak the same language (s) as the speech-language pathologist. Am. J. Speech Lang. Pathol. 26(3), 691–708 (2017)
Nelson, T.L., Mok, Z., Eecen, K.T.: Use of transcription when assessing children’s speech: Australian speech-language pathologists’ practices, challenges, and facilitators. Folia Phoniatr. Logop. 72(2), 131–142 (2020)
Oschshorn, R., Hawkins, M.: Gentle (2017)
von Platen, P.: Fine-Tune Wav2Vec2 for English ASR in Hugging Face with huggingface Transformers (2021). https://huggingface.co/blog/fine-tune-wav2vec2-english
Rosenfelder, I., et al.: Fave (forced alignment and vowel extraction) suite version 1.1.3 (2014). https://doi.org/10.5281/zenodo.9846
Schuster, M., Paliwal, K.: Bidirectional recurrent neural networks. IEEE Trans. Signal Process. 45(11), 2673–2681 (1997)
Zhu, J., Zhang, C., Jurgens, D.: Phone-to-audio alignment without text: a semi-supervised approach. In: Proceedings of ICASSP, pp. 8167–8171. IEEE (2022)
Acknowledgement
This work has been supported by the Western Australian Future Health Research and Innovation Fund, which is an initiative of the WA State Government”. This work is being conducted to inform a larger research program, being led by a team of researchers at Curtin university. The research program is focused on the development of an application that will provide objective kinematic and acoustic measurements, to support speech language pathologists in the diagnosis of speech sound disorders. The authors would also like to thank Pawsey supercomputing centre for their support.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Wohlan, B., Pham, DS., Chan, K.Y., Ward, R. (2022). A Text-Independent Forced Alignment Method for Automatic Phoneme Segmentation. In: Aziz, H., Corrêa, D., French, T. (eds) AI 2022: Advances in Artificial Intelligence. AI 2022. Lecture Notes in Computer Science(), vol 13728. Springer, Cham. https://doi.org/10.1007/978-3-031-22695-3_41
Download citation
DOI: https://doi.org/10.1007/978-3-031-22695-3_41
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-22694-6
Online ISBN: 978-3-031-22695-3
eBook Packages: Computer ScienceComputer Science (R0)