Summary
Objective
Doctors, nowadays, primarily use auditory-perceptual evaluation, such as the grade,
roughness, breathiness, asthenia, and strain scale, to evaluate voice quality and
determine the treatment. However, the results predicted by individual physicians often
differ, because of subjective perceptions, and diagnosis time interval, if the patient's
symptoms are hard to judge. Therefore, an accurate computerized pathological voice
quality assessment system will improve the quality of assessment.
Method
This study proposes a self_attention-based system, with a deep learning technology,
named self_attention-based bidirectional long-short term memory (SA BiLSTM). Different
pitches [low, normal, high], and vowels [/a/, /i/, /u/], were added into the proposed
model, to make it learn how professional doctors evaluate the grade, roughness, breathiness,
asthenia, and strain scale, in a high dimension view.
Results
The experimental results showed that the proposed system provided higher performance
than the baseline system. More specifically, the macro average of the F1 score, presented
as decimal, was used to compare the accuracy of classification. The (G, R, and B)
of the proposed system were (0.768±0.011, 0.820±0.009, and 0.815±0.009), which is
higher than the baseline systems: deep neural network (0.395±0.010, 0.312±0.019, 0.321±0.014)
and convolution neural network (0.421±0.052, 0.306±0.043, 0.3250±0.032) respectively.
Conclusions
The proposed system, with SA BiLSTM, pitches, and vowels, provides a more accurate
way to evaluate the voice. This will be helpful for clinical voice evaluations and
will improve patients’ benefits from voice therapy.
Key words
To read this article in full you will need to make a payment
Purchase one-time access:
Academic & Personal: 24 hour online accessCorporate R&D Professionals: 24 hour online accessOne-time access price info
- For academic or personal research use, select 'Academic and Personal'
- For corporate R&D use, select 'Corporate R&D Professionals'
Subscribe:
Subscribe to Journal of VoiceAlready a print subscriber? Claim online access
Already an online subscriber? Sign in
Register: Create an account
Institutional Access: Sign in to ScienceDirect
References
- Clinical voice disorders: an interdisciplinary approach.Mayo Clinic Proceedings. 66. Elsevier, 1991: 656
- Voice disorders in the general population: prevalence, risk factors, and occupational impact.Laryngoscope. 2005; 115: 1988-1995
- Auditory-perceptual evaluation of disordered voice quality.Folia Phoniatr Logop. 2009; 61: 49-56
- GRBAS-scaling of pathological voices: reliability, clinical relevance, and differentiated correlation with acoustic measurements, especially with cepstral measurements.in: Proceedings of the 22th World Congress IALP. 1992
- Reliability in perceptual analysis of voice quality.J Voice. 2005; 19: 555-573
- Test-retest study of the GRBAS scale: influence of experience and professional background on perceptual rating of voice quality.J Voice. 1997; 11: 74-80
- Modulation spectra morphological parameters: a new method to assess voice pathologies according to the GRBAS scale.BioMed Res Int. 2015;
- Reliability of perceptions of voice quality: evidence from a problem asthma clinic population.J Laryngol Otol. 2009; 123: 755-763
- Age and changes in vocal jitter.J Gerontol. 1980; 35: 194-198
- Reliable jitter and shimmer measurements in voice clinics: the relevance of vowel, gender, vocal intensity, and fundamental frequency effects in a typical clinical task.J Voice. 2011; 25: 44-53
- Vocal acoustic analysis–jitter, shimmer and hnr parameters.Procedia Technol. 2013; 9: 1112-1122
- Comparing reliability of perceptual ratings of roughness and acoustic measures of jitter.J Speech Lang Hear Res. 1995; 38: 26-32
- Deep learning.Nature. 2015; 521: 436-444
- Detection of pathological voice using cepstrum vectors: a deep learning approach.J Voice. 2019; 33: 634-641
- Psycho-acoustic evaluation of voice.Clini Exam Voice. 1981; : 81-84
Hidaka S, Lee Y, Wakamiya K, et al. Automatic Estimation of Pathological Voice Quality Based on Recurrent Neural Network Using Amplitude and Phase Spectrogram. In INTERSPEECH. 2020:3880-3884.
- Objective assessment of pathological voice using artificial intelligence based on the GRBAS scale.J Voice. 2021; https://doi.org/10.1016/j.jvoice.2021.11.021
- Multimodal and multi-output deep learning architectures for the automatic assessment of voice quality using the GRB scale.IEEE J Selec Top Signal Process. 2019; 14: 413-422
García MA, Rosset AL. Deep Neural Network for Automatic Assessment of Dysphonia. arXiv preprint arXiv:2202.12957. 2022.
- Classification of voice disorders using a one-dimensional convolutional neural network.J Voice. 2022; 36: 15-20
- Vowel-and text-based cepstral analysis of chronic hoarseness.J Voice. 2012; 26: 416-424
- Validity, reliability and reproducibility of the “extended GRBAS scale,” a comprehensive perceptual evaluation of dysphonia.J Voice. 2022;
- Perceptual evaluation of voice quality and its correlation with acoustic measurements.J Voice. 2004; 18: 299-304
- Acoustic parameters for classification of breathiness in continuous speech according to the GRBAS scale.J Voice. 2014; 28 (653-e9)
- Perceptual and quantitative assessment of dysphonia across vowel categories.J Voice. 2019; 33: 473-481
- Examining relationships between GRBAS ratings and acoustic, aerodynamic and patient-reported voice measures in adults with voice disorders.J Voice. 2021;
- Effect of endoscopic glottoplasty on acoustic measures and quality of voice: a systematic review and meta-analysis.J Voice. 2020;
- Covariation between voice quality and pitch: revisiting the case of Mandarin creaky voice.J Acoust Soc Am. 2017; 142: 1693-1706
- Throaty voice quality: subglottal pressure, voice source, and formant characteristics.J Voice. 2006; 20: 25-37
- Attention is all you need.Advan Neural Inform Process Syst. 2017; : 30
- Overview of the transformer-based models for NLP Tasks.in: 15th IEEE Conference on Computer Science and Information Systems (FedCSIS). 2020: 179-183
Woldert-Jokisz B. Saarbruecken voice database. 2007.
- Computing. Estimation of prediction error by using K-fold cross-validation.Statist Comput. 2011; 21: 137-146
- Analysis of k-fold cross-validation over hold-out validation on colossal datasets for quality classification.in: IEEE 6th International conference on advanced computing (IACC). 2016
- Comparative analysis of CNN and RNN for voice pathology detection.Biomed Res Int. 2021; : 2021
- Performance comparison of heterogeneous classifiers for detection of Parkinson's disease using voice disorder (dysphonia).in: International Conference on Informatics, Electronics & Vision (ICIEV). 2014
- Classification of heart sound signals using a novel deep WaveNet model.Comput Met Prog Biom. 2020; 196105604
- A speech enhancement approach using piecewise linear approximation of an explicit model of environmental distortions.in: Ninth annual conference of the international speech communication association (INTERSPEECH). 2008
- Context-aware self-attention networks for natural language processing.Neurocomputing. 2021; 458: 157-169
- State-of-the-art speech recognition using multi-stream self-attention with dilated 1d convolutions.IEEE Automatic Speech Recognition and Understanding Workshop (ASRU). 2019: 54-61
- Activation functions in neural networks.Towards Data Sci. 2017; 6: 310-316
Agarap AF. Deep learning using rectified linear units (relu). arXiv preprint arXiv: 1803.08375. 2018.
- Generalized cross entropy loss for training deep neural networks with noisy labels.Advan Neural Inform Process Syst. 2018; : 31
- Can cross entropy loss be robust to label noise?.in: Proceedings of the Twenty-Ninth International Conference on International Joint Conferences on Artificial Intelligence (IJCAI). 2021: 2206-2212
- Incorporating nesterov momentum into adam.in: International Conference on Learning Representations (ICLR). 2016
- Pitch-synchronous waveform processing techniques for text-to-speech synthesis using diphones.Speech Comm. 1990; 9: 453-467
- MBR-PSOLA: text-to-speech synthesis based on an MBE re-synthesis of the segments database.Speech Comm. 1993; 13: 435-440
- Audio augmentation for speech recognition.in: Sixteenth Annual Conference Of The International Speech Communication Association (INTERSPEECH). 2015
- An objective evaluation framework for pathological speech synthesis.in: Speech Communication; 14th ITG Conference. 2021
- Scikit-learn: machine learning in python.J Mach Learn Res. 2011; 12: 2825-2830
Powers DM. Evaluation: from precision, recall and F-measure to ROC, informedness, markedness and correlation. arXiv preprint arXiv:2010.16061. 2020.
Gowda T, You W, Lignos C, et al. Macro-average: rare types are important too. arXiv preprint arXiv:2104.05700. 2021.
- The role of voice therapy in male-to-female transsexuals.Current Opinion Otolaryngo Head Neck Surgery. 2002; 10: 173-177
- Visualizing data using t-SNE.J Mach Learn Res. 2008; 9
- Improved environment aware based noise reduction system for cochlear implant users based on a knowledge transfer approach: development and usability study.J Med Int Res. 2021; 23: e25460
Article info
Publication history
Published online: January 31, 2023
Accepted:
December 28,
2022
Publication stage
In Press Corrected ProofFootnotes
This study was supported by the industry-academia cooperation project of APrevent Medical Inc. (109J052), the National Science and Technology Council of Taiwan (MOST 110-2218-E-A4S9A-501) and (MOST 111-2221-E-A49-041-MY2).
Identification
Copyright
© 2023 The Voice Foundation. Published by Elsevier Inc. All rights reserved.
