Voice detection and pattern recognition using neck skin vibration signals
Notice bibliographique
Résumé
During human phonation, vibrations of the human vocal folds generate glottal periodic acoustic waves, which propagate through the vocal tract and are radiated from the nose and lips. These waves are also structurally transmitted through the bodies to reach different regions of the body. Long-term human voice recording is frequenctly used to facilitate the clinical diagnosis of chronic voice problems. Many devices, such as the microphone, have been criticized for speech privacy disclosure and noise susceptibility, which make long-term voice monitoring difficult to implement. In this dissertation, a portable neck surface accelerometer (NSA) was designed and fabricated to monitor long-term voice use and study the voice features under different vocal conditions, using both subjective and objective voice assessment methods.To investigate the robustness of glottal inverse filtering algorithms, mathematical models of sound propagation in the human respiratory system based on lumped analog circuits were used. The modeling uncertainties of glottal inverse filtering methods were then investigated using synthesized and recorded voice data for different vowels and voice types. The results showed that the accuracies of the supraglottal inverse filtering method varies notably with vowel type and voice type. Consequently, inverse filtering methods are not particularly advantageous for voice monitoring.To investigate the automatic recognition of different voice types (modal, breathy, and pressed), voice data for different voice types were collected from 31 native Canadian English speakers for single vowel phonation using the NSA and the microphone simultaneously. Auditory-perceptual ratings were conducted by five clinically certified speech language pathologists to categorize voice types using the microphone recordings. Congruent NSA samples were analyzed to find trends in extracted voice metrics, such as spectral harmonics, spectral entropy, jitter, and shimmer. An overall classification accuracy greater than 80% was achieved using supervised learning techniques. To investigate the variations of voice quality induced by intensive voice use, a dose-based vocal loading task (VLT) experiment was conducted on nine native Canadian English speakers. The experimental protocol included six successive dose-calibrated VLT sessions followed by a rest session. Voice qualities were evaluated using two standard subjective voice quality assessment methods (the CAPE-V and the SAVRa) at eight time points during this experiment. Results showed that the CAPE-V and the SAVRa ratings consistently followed similar trends. Across-session variations of four microphone features (fundamental frequency, SPL, duty ratio, shimmer) and two NSA features (shimmer, spectral tilt) were closely correlated (rmin>0.7) with each other and showed a general trend of a vocal adjustment period followed by a vocal saturation period. This trend was also consistent with that of the subjective ratings. Overall, the work reported in this thesis supports the novel concept of continuous ambulatory voice monitoring for the detection of vocal fatigue using filted data and machine learning algorithms. The main limitation of this work is the use of signle vowel recordings as opposed to running speech, which will be addressed in future work
Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.
Comment cette classification a été obtenuedéplier
Prédiction distillée sur la base complète
Imitation des enseignantsNi prévalence calibrée, ni vérité terrain. Validation humaine à venir. Apprise à partir de 10 348 étiquettes directes de Codex et de 10 348 étiquettes directes de Gemma. Le mode candidate est l'union des têtes enseignantes seuillées; le consensus est leur intersection. Ces sorties portent le statut machine_predicted_unvalidated et ne sont ni des étiquettes humaines ni des étiquettes directes de modèles de pointe.
Scores Codex et Gemma par catégorie
| Catégorie | Codex | Gemma |
|---|---|---|
| Métarecherche | 0,000 | 0,000 |
| Méta-épidémiologie (sens strict) | 0,000 | 0,000 |
| Méta-épidémiologie (sens large) | 0,000 | 0,000 |
| Bibliométrie | 0,000 | 0,000 |
| Études des sciences et des technologies | 0,000 | 0,000 |
| Communication savante | 0,000 | 0,000 |
| Science ouverte | 0,000 | 0,000 |
| Intégrité de la recherche | 0,000 | 0,000 |
| Charge utile insuffisante (le modèle a refusé de juger) | 0,000 | 0,000 |
Scores machine (provisoires)
Les deux têtes enseignantes du modèle étudiant, lues sur ce travail. Un score ordonne la base pour la relecture; il n'affirme jamais une catégorie, et le statut de validation accompagne chaque rangée tel quel.
Scores de référence d'un modèle non mature (critères de maturité non atteints, 7 itérations). Un score ordonne; il n'affirme jamais une catégorie.
score_only:v0-immature-baseline · tel quel depuis la passe de notation : score_only signifie que le nombre peut ordonner les travaux, et qu'aucune étiquette de catégorie n'en découleClassification
machine, non validéePrédiction automatique; un appel candidat d’une seule tête enseignante, pas un consensus.
Le détail, modèle par modèle et score par score, se trouve en fin de page sous « Comment cette classification a été obtenue ».