Wyniki wyszukiwania - BazTech

Ograniczanie wyników

2 Vibrations in Physical Systems

Znaleziono wyników: 2

Liczba wyników na stronie

Wyniki wyszukiwania

Sortuj według:

Ogranicz wyniki do:

GFCC-based x-vectors for Reinke’s edema detection

Kotarba Katarzyna

Vibrations in Physical Systems

2022

Vol. 33, nr 3

art. no. 2022307

Automatic assessment of voice disorders is one of the most important applications of speech signal analysis. Various algorithms utilizing both sustained vowels and continuous speech have been successfully used to perform detection of many voice pathologies, e.g. dysphonia, laryngitis, and vocal folds paralysis. However, algorithms described in literature used for classification of Reinke’s edema - one of the most severe smoking-induced voice conditions - are scarce and rely mostly on speech signals containing sustained vowels. In this paper, a method incorporating gammatone frequency cepstral coefficients (GFCC) based x-vectors extracted from continuous speech is presented. The extracted x-vectors are used to train a SGD classifier performing Reinke’s edema detection. For validation folds, the proposed method yielded AUC ROC, accuracy, recall, and specificity of 0.96 (±0.03), 0.94 (±0.02), 0.92 (±0.03), and 0.94 (±0.02), respectively. For testing set, the method yielded AUC ROC, accuracy, recall, and specificity of 0.98, 0.89, 0.88, and 0.89, respectively.

Voice pathology assessment using x-vectors approach

Kotarba Katarzyna, Kotarba Michał

Vibrations in Physical Systems

2021

Vol. 32, nr 1

art. no. 2021108

Voice pathology assessment using sustained vowels has proven to be effective and reliable. However, only a few studies regarding detection of pathological speech based on continuous speech are available. In this study we evaluate the usefulness of various regression models trained on continuous speech recordings from Saarbruecken Voice Database in the detection of voice pathologies. The recordings were used for extraction of speaker embeddings called x-vectors based on mel-frequency cepstral coefficients and gammatone frequency cepstral coefficients. Since the dataset used in this study is imbalanced, various over- and undersampling techniques were applied to the training set to ensure robustness of models’ decision boundaries. The models were trained on both imbalanced and resampled training sets using 5-fold cross-validation. The best results were obtained for Multi Layer Perceptron trained on GFCC-based x-vectors, achieving accuracy of 0.8184, F1-score of 0.8212, and ROC AUC score of 0.8810 for the testing set.