Multiple Vowels Repair Based on Pitch Extraction and Line Spectrum Pair Feature for Voice Disorder

Zhang, Tao; Shao, Yangyang; Wu, Yaqin; Pang, Zhibo; Liu, Ganjun

doi:10.1109/jbhi.2020.2978103

Cited by 17 publications

(4 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Only a few researchers, such as Guedes et al [28], have reported an accuracy of 80% employing CNNs. Zhang et al [29] used a deep neural network (DNN) model as well, however, the findings were not adequately documented. We did a thorough review of the literature to identify the most effective characteristics and classifiers for developing a novel system for speech pathology detection.…”

Section: Literature Surveymentioning

confidence: 99%

“…Standard methods, such as the median magnitude differential approach and the self-correlating system at maximum, which cause half and double-half flaws are susceptible to change during the tonnage removal process, The cepstrum system can approximate by separating the acoustic pulse cepstrum with the vocal cord cepstrum, the pitch can be determined. It has a good detection ability for simple voice signals but not for complex measures [29]. CNN Architecture: The routing layers and grouping layers that makeup CNN's hierarchy levels are defined by a very large amount of charts.…”

Section: Literature Surveymentioning

confidence: 99%

See 1 more Smart Citation

A Comparative Analysis of Convolutional Neural Network and Recurrent Neural Network for Voice Disease Detection

2023

IRJMETS

View full text Add to dashboard Cite

Early diagnosis, monitoring, and effective diagnostics of abnormal speech can greatly benefit from computerized acoustic examination. This study presents a methodology for identifying voice pathology by employing Singular Value Decomposition (SVD) datasets for feature extraction. Acoustic measurements are utilized to assess the health of the voice, and the accuracy of these parameters influences speech noise detection methods. The extracted features are then processed through a system consisting of 27 neural network layers, comprising Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN). A 10-fold validation technique is employed to split the dataset into training and testing datasets, enabling the evaluation of the system's performance. Experimental results demonstrate an observed accuracy of 87.11% for the CNN model and 86.52% for the RNN model. The classifier's performance is further assessed using 10-fold cross-validation. The implementation of the system's program is carried out in Python, utilizing the TensorFlow package, and computations are executed on a single NVidia Titan X GPU. The Linux operating system is employed for all computations. The proposed methodology offers promising potential in the early identification and monitoring of voice pathologies, providing a valuable tool for healthcare professionals in the field of speech analysis and diagnostics.

show abstract

Section: Literature Surveymentioning

confidence: 99%

Section: Literature Surveymentioning

confidence: 99%

A Comparative Analysis of Convolutional Neural Network and Recurrent Neural Network for Voice Disease Detection

2023

IRJMETS

View full text Add to dashboard Cite

show abstract

“…Also, there is not much work done for voice pathology using a convolutional neural network. Only Guedes et al [18] designed a system and reported an accuracy of 80%, and Zhang et al [19] also use the DNN model which was machine learning where outcomes were missing. So after a detailed literature review, it was concluded that a novel system can be proposed using pitch, 13 MFCC, rolloff, ZCR, energy entropy, spectral flux, spectral centroid, and energy as features and RNN as a classifier ).…”

Section: Related Workmentioning

confidence: 99%

“…By distinguishing the acoustic pulse cepstrum from the vocal tract cepstrum, the cepstrum system may approximate the pitch. At the cost of complex measurements, it has high detection performance for regular voice signal [19].…”

Section: Mel-frequency Cepstral Coefficients (Mfcc)mentioning

confidence: 99%

Comparative Analysis of CNN and RNN for Voice Pathology Detection

Syed

Rashid

Hussain

et al. 2021

BioMed Research International

View full text Add to dashboard Cite

Diagnosis on the basis of a computerized acoustic examination may play an incredibly important role in early diagnosis and in monitoring and even improving effective pathological speech diagnostics. Various acoustic metrics test the health of the voice. The precision of these parameters also has to do with algorithms for the detection of speech noise. The idea is to detect the disease pathology from the voice. First, we apply the feature extraction on the SVD dataset. After the feature extraction, the system input goes into the 27 neuronal layer neural networks that are convolutional and recurrent neural network. We divided the dataset into training and testing, and after 10 k-fold validation, the reported accuracies of CNN and RNN are 87.11% and 86.52%, respectively. A 10-fold cross-validation is used to evaluate the performance of the classifier. On a Linux workstation with one NVidia Titan X GPU, program code was written in Python using the TensorFlow package.

show abstract

Significance of Acoustic Features in Vertebrate Vocalisations Related Applications

Ramashini

Gamage

Murugiah³

et al. 2021

Acoustic-Based Applications for Vertebrate Vocalization

View full text Add to dashboard Cite

Multiple Vowels Repair Based on Pitch Extraction and Line Spectrum Pair Feature for Voice Disorder

Cited by 17 publications

References 20 publications

A Comparative Analysis of Convolutional Neural Network and Recurrent Neural Network for Voice Disease Detection

A Comparative Analysis of Convolutional Neural Network and Recurrent Neural Network for Voice Disease Detection

Comparative Analysis of CNN and RNN for Voice Pathology Detection

Significance of Acoustic Features in Vertebrate Vocalisations Related Applications

Contact Info

Product

Resources

About