Dataset Information

Robust Vocal Quality Feature Embeddings for Dysphonic Voice Detection.

ABSTRACT: Approximately 1.2% of the world's population has impaired voice production. As a result, automatic dysphonic voice detection has attracted considerable academic and clinical interest. However, existing methods for automated voice assessment often fail to generalize outside the training conditions or to other related applications. In this paper, we propose a deep learning framework for generating acoustic feature embeddings sensitive to vocal quality and robust across different corpora. A contrastive loss is combined with a classification loss to train our deep learning model jointly. Data warping methods are used on input voice samples to improve the robustness of our method. Empirical results demonstrate that our method not only achieves high in-corpus and cross-corpus classification accuracy but also generates good embeddings sensitive to voice quality and robust across different corpora. We also compare our results against three baseline methods on clean and three variations of deteriorated in-corpus and cross-corpus datasets and demonstrate that the proposed model consistently outperforms the baseline methods.

SUBMITTER: Zhang J

PROVIDER: S-EPMC10602198 | biostudies-literature | 2023

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Robust Vocal Quality Feature Embeddings for Dysphonic Voice Detection.

Zhang Jianwei J Liss Julie J Jayasuriya Suren S Berisha Visar V

IEEE/ACM transactions on audio, speech, and language processing 20230328

Approximately 1.2% of the world's population has impaired voice production. As a result, automatic dysphonic voice detection has attracted considerable academic and clinical interest. However, existing methods for automated voice assessment often fail to generalize outside the training conditions or to other related applications. In this paper, we propose a deep learning framework for generating acoustic feature embeddings sensitive to vocal quality and robust across different corpora. A contras ...[more]

PMID: 37899766

Similar Datasets

Project description:ImportanceEssential vocal tremor (EVT) is a neurologic voice disorder characterized by periodic fluctuations in pitch and loudness that can hinder intelligibility. Defining the normative range of vocal tremor frequency may assist in diagnosis and provide insight into disease mechanisms.ObjectiveTo characterize the normative voice tremor frequency in EVT (in hertz).Design, setting, and participantsCross-sectional observational study of 160 patients with EVT. The setting was a tertiary voice center. Participants were identified from a database of consecutive patients diagnosed as having laryngeal movement disorders between January 1, 1990, and April 1, 2017.Main outcomes and measuresThe following 3 methods measured the frequency of tremor experienced by patients with EVT: perceptual method, computerized peak detection method, and laryngeal electromyography method. Within-person and population-level tremor frequencies were compared across modalities to assess measurement reliability and consistency and to characterize the normal distribution of tremor frequencies in this population.ResultsAmong 160 participants (median age, 70 years; interquartile range [IQR], 64-77 years; 90.6% female [n = 145]), the median frequency of EVT was consistently between 4 and 5 Hz across all 3 methods (perceptual, 4.8 Hz [IQR, 4.4-5.5 Hz]; computerized peak detection, 4.6 Hz [IQR, 4.2-5.0 Hz]; and laryngeal electromyography, 4.3 Hz [IQR, 3.8-5.0 Hz]). The mean in-person differences between each measurement method were not clinically meaningful (range, 0.1-0.5 Hz). Including all interquartile ranges across measurement modalities, the normative tremor frequency range for EVT was 3.8 to 5.5 Hz.Conclusions and relevanceTo our knowledge, this is the largest study to date to characterize the normal frequency of tremor in patients with EVT. The normative frequency of EVT (range, 3.8-5.5 Hz) falls within a much narrower range than previously reported. Those whose frequency is outside this range may still have EVT but should be carefully evaluated for potential other causes of vocal tremor. Defining characteristics of EVT may aid appropriate diagnosis and improve understanding of this disease.

Dataset Information

Robust Vocal Quality Feature Embeddings for Dysphonic Voice Detection.

Publications

Robust Vocal Quality Feature Embeddings for Dysphonic Voice Detection.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets