A Deep‐Learning Model for Multi‐class Audio Classification of Vocal Fold Pathologies in Office Stroboscopy
Open accessAbstract
Objective: To develop and validate a deep‐learning classifier trained on voice data extracted from videolaryngostroboscopy recordings, differentiating between three different vocal fold (VF) states: healthy (HVF), unilateral paralysis (UVFP), and VF lesions, including benign and malignant pathologies. Methods: Patients with UVFP ( n = 105), VF lesions ( n = 63), and HVF ( n = 41) were retrospectively identified. Voice samples were extracted from stroboscopic videos (Pentax Laryngeal Strobe Model 9400), including sustained /i/ phonation, pitch glide, and /i/ sniff task. Extracted audio files were converted into Mel‐spectrograms. Voice samples were independently divided into training (80%), validation (10%), and test (10%) by patient. Pretrained ResNet18 models were trained to classify (1) HVF and pathological VF (lesions and UVFP), and (2) HVF, UVFP, and VF lesions. Both classifiers were further validated on an external dataset consisting of 12 UVFP, 13 VF lesions, and 15 HVF patients. Model performances were evaluated by accuracy and F1‐score. Results: When evaluated on a hold‐out test set, the binary classifier demonstrated stronger performance compared to the multi‐class classifier (accuracy 83% vs. 40%; F1‐score 0.90 vs. 0.36). When evaluated on an external dataset, the binary classifier achieved an accuracy of 63% and F1‐score of 0.48, compared to 35% and 0.25 for the multi‐class classifier. Conclusions: Deep‐learning classifiers differentiating HVF, UVFP, and VF lesions were developed using voice data from stroboscopic videos. Although healthy and pathological voice were differentiated with moderate accuracy, multi‐class classification lowered model performance. The model performed poorly on an external dataset. Voice captured in stroboscopic videos may have limited diagnostic value, though further studies are needed. Level of Evidence: 4 Laryngoscope , 135:2428–2436, 2025
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
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?
Laryngeal imaging is the reference standard for laryngeal evaluation, but access depends on practitioner expertise, equipment availability, and cost, and neuromuscular disorders may present without visible structural lesions.
“Laryngeal imaging—including videolaryngoscopy, stroboscopy, and high-speed videoendoscopy—represents the reference standard for laryngeal evaluation [9,16,17,18]; however, access depends on practitioner expertise, equipment availability, and cost, while neuromuscular disorders may present without visible structural lesions, complicating diagnosis [19].”
From Diagnostic Accuracy of Artificial Intelligence in Laryngeal Disorders: An Integrative Review (2026) · cites this paper as [16]
The Introduction (page not provided) calls videolaryngostroboscopy a mainstay of objective laryngeal evaluation and says access varies by region and is concentrated in specialist practices. It does not establish equipment availability or cost as access barriers, address all three named imaging methods as a group, or say that neuromuscular disorders may lack visible structural lesions. (LLM, conf=0.85)
Experimental — accuracy not yet measuredComparative evidence suggests that architectural families are better suited to specific data regimes than any universally optimal model.
“Comparative evidence suggests that architectural families are better suited to specific data regimes than any universally optimal model [16,26,56,61].”
From Diagnostic Accuracy of Artificial Intelligence in Laryngeal Disorders: An Integrative Review (2026) · cites this paper as [16]
The Results section (page unavailable) reports that an MLP performed best for binary classification, while a random forest performed best for multiclass classification. The Discussion also reports that dataset size affected performance. These findings support task- and data-dependent performance in this study, but do not establish that architectural families are generally suited to particular data regimes or rule out a universally optimal model. (LLM, conf=0.83)
Experimental — accuracy not yet measuredIn some cases, accuracy and sensitivity decrease by about 10 to 30 percentage points under external validation conditions.
“In some cases, accuracy and sensitivity decrease by about 10 to 30 percentage points under these conditions [16,64].”
From Diagnostic Accuracy of Artificial Intelligence in Laryngeal Disorders: An Integrative Review (2026) · cites this paper as [16]
The Results (Performance of Deep Learning Neural Networks; page unavailable) show that binary-classifier accuracy fell by 15–20 percentage points on the external dataset compared with the hold-out test set. The passage does not give a comparable numerical change in sensitivity, and some multi-class accuracy results did not decline, so the full claim is not supported by our copy. (LLM, conf=0.84)
Experimental — accuracy not yet measuredCNNs have been widely applied to vocal spectrograms and laryngoscopic images, while recurrent and bidirectional architectures capture temporal dynamics of pathological phonation.
“Convolutional neural networks (CNNs) have been widely applied to vocal spectrograms and laryngoscopic images [24,33], while recurrent and bidirectional architectures capture temporal dynamics of pathological phonation [16,56,57].”
From Diagnostic Accuracy of Artificial Intelligence in Laryngeal Disorders: An Integrative Review (2026) · cites this paper as [16]
Across the 88 included studies, only 7 studies performed both internal and external validation.
“Across the 88 included studies, only 7 studies performed both internal and external validation [12,16,26,54,64,65,69].”
From Diagnostic Accuracy of Artificial Intelligence in Laryngeal Disorders: An Integrative Review (2026) · cites this paper as [16]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · see citing works
(2008) · Journal of Machine Learning Research · Cited by 33,051 · see citing works
PyTorch: An Imperative Style, High-Performance Deep Learning Library
Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G., Killeen, T., Lin, Z., Gimelshein, N., Antiga, L., et al. (2019) · arXiv · doi:10.48550/arxiv.1912.01703 · Cited by 31,473 · see citing works
Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization
Selvaraju, R. R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., & Batra, D. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.74 · Cited by 24,513 · see citing works
librosa: Audio and Music Signal Analysis in Python
McFee, B., Raffel, C., Liang, D., Ellis, D. P. W., McVicar, M., Battenberg, E., & Nieto, O. (2015) · Proceedings of the Python in Science Conferences · doi:10.25080/majora-7b98e3ed-003 · Cited by 3,099 · see citing works
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 · see citing works
Multimodal machine learning in precision health: A scoping review
Kline, A., Wang, H., Li, Y., Dennis, S., Hutch, M., Xu, Z., Wang, F., Cheng, F., & Luo, Y. (2022) · npj Digital Medicine · doi:10.1038/s41746-022-00712-8 · Cited by 528 · cited 1× · see citing works
“34 With DL models applied to VLS, it will be important to explore whether combining audio with video data enhances performance compared to using video input alone.”
Ramzan, F., Khan, M. U. G., Rehmat, A., Iqbal, S., Saba, T., Rehman, A. U., & Mehmood, Z. (2019) · Journal of Medical Systems · doi:10.1007/s10916-019-1475-2 · Cited by 446 · see citing works
Vocal fold physiology : voice quality control
藤村, 靖., & 平野, 実. (1995) · Singular Pub. Group eBooks · Cited by 207 · see citing works
Differences in voice quality between men and women: Use of the long-term average spectrum (LTAS)
Mendoza, E., Naranjo, N. V., Muñoz, J., & Trujillo, H. M. (1996) · Journal of Voice · doi:10.1016/s0892-1997(96)80019-1 · Cited by 175 · see citing works
Voice Disorder Identification by Using Machine Learning Techniques
Verde, L., De Pietro, G., & Sannino, G. (2018) · IEEE Access · doi:10.1109/access.2018.2816338 · Cited by 156 · see citing works
Data Augmentation and Deep Learning Methods in Sound Classification: A Systematic Review
Abayomi‐Alli, O., Damaševičius, R., Qazi, A., Adedoyin-Olowe, M., & Misra, S. (2022) · Electronics · doi:10.3390/electronics11223795 · Cited by 151 · see citing works
Automatic Recognition of Laryngoscopic Images Using a Deep‐Learning Technique
Ren, J., Jing, X., Wang, J., Ren, X., Yang, X., Yang, Q., Ma, L., Sun, Y., Xu, W. H., Yang, N., et al. (2020) · The Laryngoscope · doi:10.1002/lary.28539 · Cited by 135 · see citing works
Voice Pathology Detection and Discrimination Based on Modulation Spectral Features
Markaki, M., & Stylianou, Y. (2011) · IEEE Transactions on Audio Speech and Language Processing · doi:10.1109/tasl.2010.2104141 · Cited by 131 · see citing works
Altman, K. W. (2007) · Otolaryngologic Clinics of North America · doi:10.1016/j.otc.2007.05.011 · Cited by 127 · see citing works
Objective Voice Quality Analysis Before and After Onset of Unilateral Vocal Fold Paralysis
Hartl, D. M., Hans, S., Vaissière, J., Riquet, M., & Brasnu, D. F. (2001) · Journal of Voice · doi:10.1016/s0892-1997(01)00037-6 · Cited by 91 · see citing works
Hu, H.-C., Chang, S.-Y., Wang, C.-H., Li, K.-J., Cho, H.-Y., Chen, Y.-T., Lu, C.-J., Tsai, T.-P., & Lee, O. K.-S. (2021) · Journal of Medical Internet Research · doi:10.2196/25247 · Cited by 82 · see citing works
Voice disorder classification using convolutional neural network based on deep transfer learning
Peng, X., Xu, H., Liu, J., Wang, J., & He, C. (2023) · Scientific Reports · doi:10.1038/s41598-023-34461-9 · Cited by 65 · see citing works
Convolutional Neural Networks for Pathological Voice Detection
Wu, H., Soraghan, J. J., Lowit, A., & Caterina, G. D. (2018) · Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc.2018.8513222 · Cited by 63 · see citing works
Cho, W. K., & Choi, S.-H. (2020) · Journal of Voice · doi:10.1016/j.jvoice.2020.08.003 · Cited by 47 · see citing works
Applications of Artificial Intelligence to Office Laryngoscopy: A Scoping Review
Yao, P., Usman, M., Chen, Y. H., German, A., Andreadis, K., Mages, K. C., & Rameau, A. (2021) · The Laryngoscope · doi:10.1002/lary.29886 · Cited by 46 · see citing works
Self-citationVoice as an AI Biomarker of Health—Introducing Audiomics
Bensoussan, Y. E., Elemento, O., & Rameau, A. (2024) · JAMA Otolaryngology–Head & Neck Surgery · doi:10.1001/jamaoto.2023.4807 · Cited by 44 · see citing works
Self-citationVideomics: bringing deep learning to diagnostic endoscopy
Paderno, A., Holsinger, F. C., & Piazza, C. (2021) · Current Opinion in Otolaryngology & Head & Neck Surgery · doi:10.1097/moo.0000000000000697 · Cited by 43 · see citing works
Kwon, I., Wang, S.-G., Shin, S.-C., Cheon, Y.-I., Lee, B.-J., Lee, J.-C., Lim, D.-W., Jo, C., Cho, Y., & Shin, B.-J. (2025) · Journal of Voice · doi:10.1016/j.jvoice.2022.07.007 · Cited by 43 · see citing works
Ünalan, S., Günay, O., Akkurt, İ., Günoğlu, K., & Tekin, H. O. (2024) · Journal of Radiation Research and Applied Sciences · doi:10.1016/j.jrras.2024.101080 · Cited by 42 · see citing works
Automatic classification of informative laryngoscopic images using deep learning
Yao, P., Witte, D., Gimonet, H., German, A., Andreadis, K., Cheng, M., Sulica, L., Elemento, O., Barnes, J., & Rameau, A. (2022) · Laryngoscope Investigative Otolaryngology · doi:10.1002/lio2.754 · Cited by 40 · see citing works
Self-citationUsing SincNet for Learning Pathological Voice Disorders
Hung, C.-H., Wang, S.-S., Wang, C.-T., & Fang, S.-H. (2022) · Sensors · doi:10.3390/s22176634 · Cited by 39 · see citing works
Voice-Indistinguishability: Protecting Voiceprint In Privacy-Preserving Speech Data Release
Han, Y., Li, S., Cao, Y., Ma, Q., & Yoshikawa, M. (2020) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme46284.2020.9102875 · Cited by 37 · see citing works
Development of a machine-learning based voice disorder screening tool
Reid, J., Parmar, P., Lund, T., Aalto, D. K., & Jeffery, C. C. (2022) · American Journal of Otolaryngology · doi:10.1016/j.amjoto.2021.103327 · Cited by 35 · see citing works
Malekroodi, H. S., Madusanka, N., Lee, B.-I., & Yi, M. (2024) · Bioengineering · doi:10.3390/bioengineering11030295 · Cited by 34 · see citing works
Chen, Z., Zhu, P., Qiu, W., Guo, J., & Li, Y. (2023) · International Journal of Language & Communication Disorders · doi:10.1111/1460-6984.12783 · Cited by 31 · see citing works
Decoding phonation with artificial intelligence (D e P AI): Proof of concept
Powell, M., Rodriguez-Cancio, M., Young, D., Nock, W., Abdelmessih, B. A., Zeller, A., Morales, I. P., Zhang, P., Garrett, C. G., Schmidt, D. C., et al. (2019) · Laryngoscope Investigative Otolaryngology · doi:10.1002/lio2.259 · Cited by 27 · see citing works
Combining acoustic signals and medical records to improve pathological voice classification
Fang, S., Wang, C., Chen, J., Tsao, Y., & Lin, F. (2019) · APSIPA Transactions on Signal and Information Processing · doi:10.1017/atsip.2019.7 · Cited by 26 · see citing works
Liu, G. S., Jovanovic, N., Sung, C. K., & Doyle, P. C. (2024) · Otolaryngology · doi:10.1002/ohn.809 · Cited by 26 · see citing works
Support of deep learning to classify vocal fold images in flexible laryngoscopy
Tran, B. A., Dao, T. T. P., Dung, H. D. Q., Van, B. N., Ha, C. C., Pham, N. H., Nguyen, T. C. H. T. N. C., Nguyen, T.-C., Pham, M.-K., Tran, M.-K., et al. (2023) · American Journal of Otolaryngology · doi:10.1016/j.amjoto.2023.103800 · Cited by 19 · see citing works
Trust in Machine Learning Driven Clinical Decision Support Tools Among Otolaryngologists
Chen, H. Y., Ma, X., Rives, H., Serpedin, A., Yao, P., & Rameau, A. (2024) · The Laryngoscope · doi:10.1002/lary.31260 · Cited by 18 · see citing works
Self-citationLow, D. M., Rao, V., Randolph, G., Song, P. C., & Ghosh, S. S. (2024) · PLOS Digital Health · doi:10.1371/journal.pdig.0000516 · Cited by 16 · see citing works
Seedat, N., Aharonson, V., & Hamzany, Y. (2020) · IEEE Asia-Pacific Conference on Computer Science and Data Engineering (CSDE) · doi:10.1109/csde50874.2020.9411529 · Cited by 15 · see citing works
Sound Spectrographic Analysis of the Voice of Patients with Recurrent Laryngeal Nerve Paralysis
Kim, K. M., Kakita, Y., & Hirano, M. (2009) · Folia Phoniatrica et Logopaedica · doi:10.1159/000265641 · Cited by 15 · see citing works
The Effect of Noise on Deep Learning for Classification of Pathological Voice
Hasebe, K., Fujimura, S., Kojima, T., Tamura, K., Kawai, Y., Kishimoto, Y., & Omori, K. (2024) · The Laryngoscope · doi:10.1002/lary.31303 · Cited by 11 · see citing works
The Geographic Distribution of Videolaryngostroboscopy in the United States
Davis, R. J., Exilus, S., Best, S. R., Willink, A., & Akst, L. M. (2021) · Journal of Voice · doi:10.1016/j.jvoice.2021.05.002 · Cited by 7 · see citing works
Diabetic Retinopathy Detection with Weighted Cross-Entropy Loss
Huang, J., Wu, X., Qi, H., Cheng, J.-S., & Zhang, T. (2023) · Chinese Control Conference (CCC) · doi:10.23919/ccc58697.2023.10240083 · Cited by 2 · see citing works
Welch, B., Gold, K., Gartner-Schmidt, J., et al. (2023) · doi:10.31219/osf.io/xgr2h · Cited by 1 · see citing works
Welch, B., Gold, K., Gartner‐Schmidt, J. L., Petrov, A. A., Law, A., & Helou, L. B. (2024) · Journal of Voice · doi:10.1016/j.jvoice.2023.12.007 · Cited by 1 · see citing works
Flagged first, then by citations
Research works printed but not linked: 2
- [27]Loshchilov I, Hutter F. Decoupled weight decay regularization. arXiv preprint arXiv:171105101 2017.
- [31]Lundberg S. A unified approach to interpreting model predictions. arXiv preprint arXiv:170507874 2017.
Cited by
Sorted by citations
Multimodal laryngoscopic video analysis for assisted diagnosis of vocal fold paralysis
Zhang, Y., Zou, X., Yang, J., Chen, W., Liu, J., Liang, F., & Li, M. (2025) · Computer Speech & Language · doi:10.1016/j.csl.2025.101891 · Cited by 3 · see citing works
A multimodal approach for deep-learning classification of vocal fold pathologies in stroboscopy
Surapaneni, S., Kutler, R. B., Setzen, S. A., Kim, Y. E., Yao, P., Siddiqui, S. H., Pitman, M. J., Sulica, L., Elemento, O., Khosravi, P., et al. (2026) · The Laryngoscope · doi:10.1002/lary.70355 · Cited by 2 · see citing works
Self-citationDecoding Gender in Cough Sounds: A Transformer‐Based Analysis
He, L., Li, H., Wang, S., Baik, E., Kervin, S. R., Zhao, R., Ramos, J. M., Colonel, J. T., & Rameau, A. (2026) · The Laryngoscope · doi:10.1002/lary.70393 · Cited by 2 · see citing works
Self-citationKutler, R. B., & Rameau, A. (2026) · World Journal of Otorhinolaryngology - Head and Neck Surgery · doi:10.1002/wjo2.70103 · Cited by 1 · see citing works
Self-citationDiagnostic Accuracy of Artificial Intelligence in Laryngeal Disorders: An Integrative Review
Mairesse, S., Maniaci, A., Briganti, G., & Lechien, J. R. (2026) · Journal of Personalized Medicine · doi:10.3390/jpm16060301 · Cited by 1 · see citing works
“Convolutional neural networks (CNNs) have been widely applied to vocal spectrograms and laryngoscopic images [24,33], while recurrent and bidirectional architectures capture temporal dynamics of pathological phonation [16,56,57].”
InfluentialHsiao, C., Cheng, Y.-R., Yang, C., & Hsu, F.-S. (2026) · medRxiv · doi:10.64898/2026.05.29.26354442
Beyond fingerprint and voice print: cough sequence sound for identity verification
Li, H., Wang, S., He, L., Colonel, J. T., Zhao, R., & Rameau, A. (2026) · npj Health Systems · doi:10.1038/s44401-026-00111-1
Self-citationMulti‐Classification of Vocal Cord Masses via Local‐Higher Order Graph Neural Network on Voice Data
Wang, T., Zhao, X., Li, S., Guo, H., Zheng, L., & Yu, D. (2026) · The Laryngoscope · doi:10.1002/lary.70800
Verma, H., Mishra, R., Nagamani, B., Patro, S. K., Ramavat, A. S., Bakshi, J. B., Khandnor, P., Rani, S., & Chand, T. (2026) · Medical Sciences · doi:10.3390/medsci14050581
- Authors
- Kim, Y. E., Dobko, M., Li, H., Shao, T., Periyakoil, P., Tipton, C., Colasacco, C., Serpedin, A., Elemento, O., Sabuncu, M., Pitman, M., Sulica, L., & Rameau, A.
- Affiliations
- Weill Cornell Medicine (US)
- Cornell Tech (US)
- NewYork–Presbyterian Hospital (US)
- Weill Cornell Medical College in Qatar (QA)
- Corresponding author
- Rameau, Anaïs
- ORCID
- Venue
- The Laryngoscope
- Publisher
- Wiley
- Published
- 2025-02-05
- Volume
- 135
- Issue
- 7
- Pages
- 2428-2436
- ISSN
- 0023-852X (ISSN-L)
- 1531-4995
- PMC
- PMC12234251
- OpenAlex
- W4407176216
- Cited by
- 9
- References
- 42
- Citations by year
- 2026: 8 · 2025: 1
- FWCI
- 5.74
- Type
- journal article
- Language
- en
- Open access
- bronze
- doi.org
- Locations
- The Laryngoscope (publishedVersion)
- PubMed (publishedVersion)
- Europe PMC (PubMed Central) (acceptedVersion)
- eCommons (Cornell University) (submittedVersion, cc-by-nc-nd)
- Funders
- National Institute on Aging (K76AG079040)
- Common Fund (OT2 OD032720)
- Topics
- Health Sciences › Medicine › Physiology › Voice and Speech Disorders
- Health Sciences › Medicine › Pulmonary and Respiratory Medicine › Phonocardiography and Auscultation Techniques
- Health Sciences › Health Professions › Health Information Management › Artificial Intelligence in Healthcare
- Keywords
- vocal fold pathology, videolaryngostroboscopy, unilateral vocal fold paralysis, vocal fold lesions, voice analysis, Mel spectrogram, ResNet-18, external validation
- MeSH terms
- Deep Learning, Adult, Aged, Female, Humans, Laryngeal Diseases, Laryngoscopy, Male, Middle Aged, Retrospective Studies, Video Recording, Vocal Cord Paralysis, Vocal Cords, Stroboscopy
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
