Wavelet time-frequency analysis and least squares support vector machines for the identification of voice disorders

Fonseca, Everthon Silva; Guido, Rodrigo Capobianco; Scalassara, Paulo Rogério; Maciel, Carlos Dias; Pereira, José Carlos

doi:10.1016/j.compbiomed.2006.08.008

Cited by 115 publications

(47 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The mother wavelet used in this study is reported to be effective in voice signal analysis [18][19] and is being widely used in many pathological voice analyses [17]. Due to the noise-like effect of irregularities in the vibration pattern of damaged vocal folds, the distribution manner of such variations within the whole frequency range of pathological speech signals is not clearly known.…”

Section: Mel-frequency-cepstral-coefficientsmentioning

confidence: 99%

See 1 more Smart Citation

A Hybrid of Genetic Algorithm and Support Vector Machine for Feature Reduction and Detection of Vocal Fold Pathology

Majidnezhad¹,

Kheidorov²

2013

IJIGSP

View full text Add to dashboard Cite

Abstract-Acoustic analysis is a proper method in vocal fold pathology diagnosis so that it can complement and in some cases replace the other invasive, based on direct vocal fold observation, methods. There are d ifferent approaches and algorith ms for vocal fold pathology diagnosis. These algorithms usually have three stages which are Feature Extraction, Feature Reduction and Classification. While the third stage imp lies a choice of a variety of machine learning methods (Support Vector Machines, Artificial Neural Networks, etc), the first and second stages play a critical ro le in performance and accuracy of the classification system. In this paper we present initial study of feature ext raction and feature reduction in the task of vocal fold pathology diagnosis. A new type of feature vector, based on wavelet packet decomposition and Mel-Frequency-Cepstral-Coefficients (MFCCs), is proposed. Also a new GA -based method for feature reduction stage is proposed and compared with conventional methods such as Principal Co mponent Analysis (PCA). Support vector machine is used as a classifier for evaluating the performance of the proposed method. The results show the priority of the proposed method in co mparison with the current methods.

show abstract

Section: Mel-frequency-cepstral-coefficientsmentioning

confidence: 99%

“…The most important, mult i resolution property of WPs is helpful in voice signal synthesis [16][17]. The hierarchical WP transform uses a family of wavelet functions and their associated scaling functions to decompose the original signal into subsequent sub-bands.…”

Section: Mel-frequency-cepstral-coefficientsmentioning

confidence: 99%

A Hybrid of Genetic Algorithm and Support Vector Machine for Feature Reduction and Detection of Vocal Fold Pathology

Majidnezhad¹,

Kheidorov²

2013

IJIGSP

View full text Add to dashboard Cite

show abstract

“…The outcome of such a technique is often superior to results obtained by other machine learning algorithms such as Artificial Neural Networks [8,9,10]. SVMs have been used for pattern recognition, image processing, machine learning, bioinformatics, among others [11].…”

Section: Support Vector Machinementioning

confidence: 99%

Identification of Foliar Diseases in Cotton Crop

Bernardes

Rogéri

Oliveira

et al. 2013

Lecture Notes in Computational Vision and Biomechanics

View full text Add to dashboard Cite

The pathogens manifestation in plantations are the largest cause of damage in several cultivars, which may cause increase of prices and loss of crop quality. This paper presents a method for automatic classification of cotton diseases through feature extraction of leaf symptoms from digital images. Wavelet transform energy has been used for feature extraction while Support Vector Machine has been used for classification. Five situations have been diagnosed, namely: Healthy crop, Ramularia disease, Bacterial Blight, Ascochyta Blight, and unspecified disease. Proposed MethodThe classification process was divided into two phases:  Phase 1: Finding the best feature vector for each class;  Phase 2: Create the final classification system from the best results obtained in the previous phase. Phase 1: Finding the Best Feature VectorThis phase is aimed at finding the best feature vector to represent each of the classes to be considered during classification. To achieve this goal the following steps were accomplished: Decomposition of images into multiple channels (R, G, B, H, S, V, I3a, I3b, and GL);  Application of the discrete wavelet transform (DWT) up to the third level;  Computation of the energy for each sub-band and compose the feature vector; Creation of the SVM classification environment;  Listing of the images used for training and testing;  Evaluation of the best feature vectors. Decomposition of the ImageThe decomposition of the images is the first process the system executes. In this stage an image is decomposed into nine channels, namely: R, G, B, H, S, V, I3a, I3b and GL. Application of the Discrete Wavelet TransformDiscrete Wavelet Transform (TWD) decomposition is applied up to the third level. When an image is decomposed as such it will have ten sub-bands, as illustrated in Figure 3. Note that each sub-band is identified by a number between 1 and 10. Region A1 and sub-bands 8, 9 and 10, are generated by the first level of decomposition of the DWT. Region A2 and sub-bands 5, 6 and 7 refer to the second level of decomposition, and the third level is formed by the sub-bands 1, 2, 3 and 4. Computation of the Energy for Each Sub-bandAfter applying the DWT to the three levels, the energy for each wavelet sub-band is computed. Each value obtained is inserted into a feature vector as the one illustrated in Figure 4. The vector in this figure consists of ten elements, each of them is identified by a number corresponding to the number of the sub-band in Figure 3. The energy value computed for each sub-band is stored in the corresponding vector element.Vector Features 1 2 3 4 5 6 7 8 9 10 Figure 4: Example of the structure of a feature vector. Creation of an SVM Classification EnvironmentThe network architecture used is shown in Figure 5. Note that 10 input elements are used. To each input element is assigned the value of the element of the corresponding characteristic vector. In the hidden layer there are a number of neurons (N) equal to the number of training examples, making the net converge...

show abstract

“…Some of these researches indicate that voice disorders identification can be done by the exploitation of Mel Frequency Cepstral Coefficients (MFCC) with the harmonics-to-noise ratio, normalized noise energy and glottal-to-noise excitation ratio, Gaussian mixture model was used as classifier [4]. Also, Daubechies" discrete wavelet transform, linear prediction coefficient, and least-square Support Vector Machine (LS-SVM) were investigated in [5]. In addition, a voice recognition algorithm was proposed in [6] based on the MFCC coefficients, their first and second derivatives, performance of F-ratio and Fisher"s discriminant ratio as feature reduction methods and Gaussian Mixture Model (GMM) as classifier; the main idea, here, consists in demonstrating that the detection of voice impairments can be performed using both mel cepstral vectors and their first derivative, ignoring the second derivative.…”

Section: Introductionmentioning

confidence: 99%

Artificial Neural Networks and Support Vector Machine for Voice Disorders Identification

Souissi¹,

Chérif²

2016

ijacsa

View full text Add to dashboard Cite

Abstract-The diagnosis of voice diseases through the invasive medical techniques is an efficient way but it is often uncomfortable for patients, therefore, the automatic speech recognition methods have attracted more and more interest recent years and have known a real success in the identification of voice impairments. In this context, this paper proposes a reliable algorithm for voice disorders identification based on two classification algorithms; the Artificial Neural Networks (ANN) and the Support Vector Machine (SVM). The feature extraction task is performed by the Mel Frequency Cepstral Coefficients (MFCC) and their first and second derivatives. In addition, the Linear Discriminant Analysis (LDA) is proposed as feature selection procedure in order to enhance the discriminative ability of the algorithm and minimize its complexity. The proposed voice disorders identification system is evaluated based on a widespread performance measures such as the accuracy, sensitivity, specificity, precision and Area Under Curve (AUC).

show abstract

Wavelet time-frequency analysis and least squares support vector machines for the identification of voice disorders

Cited by 115 publications

References 13 publications

A Hybrid of Genetic Algorithm and Support Vector Machine for Feature Reduction and Detection of Vocal Fold Pathology

A Hybrid of Genetic Algorithm and Support Vector Machine for Feature Reduction and Detection of Vocal Fold Pathology

Identification of Foliar Diseases in Cotton Crop

Artificial Neural Networks and Support Vector Machine for Voice Disorders Identification

Contact Info

Product

Resources

About