Academia.eduAcademia.edu

Figure 17 – uploaded by Amir Hossein Poorjam

See full PDFdownloadDownload figure
FIGURE A.1: The structure of a single neuron.

Figure 17 A.1: The structure of a single neuron.

Related Figures (34)

where 7 is the index for gradient ascent iterations, ajy is the learning rate and I is an identity matrix of dimension C. BI SSS | ee ee, Se a e The subspace matrix L is estimated over a large training dataset. The obtained subspace vectors representing the utterances in train and test datasets are used to estimate the age of speakers in this chapter. This new low-dimensional utterance representation approach was successfully applied to speaker and language/dialect recognition tasks [8].
where 7 is the index for gradient ascent iterations, ajy is the learning rate and I is an identity matrix of dimension C. BI SSS | ee ee, Se a e The subspace matrix L is estimated over a large training dataset. The obtained subspace vectors representing the utterances in train and test datasets are used to estimate the age of speakers in this chapter. This new low-dimensional utterance representation approach was successfully applied to speaker and language/dialect recognition tasks [8].
FIGURE 2.2: The age histogram of telephone speech utterances of training and testing datasets for male and female speaker.
FIGURE 2.2: The age histogram of telephone speech utterances of training and testing datasets for male and female speaker.
estimated age in year. TABLE 2.1: The results of speaker age estimation using different utterance modeling methods (the i-vector and the NFA frameworks), and different function approximation techniques (LSSVR and MLPs). CC is the Pearson correlation coefficient between actual and estimated age, and MAE is the mean-absolute error between actual and pea Sees pee pa i anes layer and 400 neurons in the second hidden layer. The preferred activation function for hidden layers was logistic sigmoid function, and in order to perform regression, a linear activation function has been utilized for the output layers. Among the various training algorithms described in Section A.2.1 of the Appendix A, the “scaled conjugate gradient” and “one step secant back-propagation” algorithms were applied for networks related to males and females, respectively. Networks were trained to minimize the mean-absolute-error between the desired and estimated outputs. To attenuate the effect of random initialization, each experiment was repeated 10 times, and the most observed result was reported. methods (the i-vector and the NFA frameworks), and different function approximation
estimated age in year. TABLE 2.1: The results of speaker age estimation using different utterance modeling methods (the i-vector and the NFA frameworks), and different function approximation techniques (LSSVR and MLPs). CC is the Pearson correlation coefficient between actual and estimated age, and MAE is the mean-absolute error between actual and pea Sees pee pa i anes layer and 400 neurons in the second hidden layer. The preferred activation function for hidden layers was logistic sigmoid function, and in order to perform regression, a linear activation function has been utilized for the output layers. Among the various training algorithms described in Section A.2.1 of the Appendix A, the “scaled conjugate gradient” and “one step secant back-propagation” algorithms were applied for networks related to males and females, respectively. Networks were trained to minimize the mean-absolute-error between the desired and estimated outputs. To attenuate the effect of random initialization, each experiment was repeated 10 times, and the most observed result was reported. methods (the i-vector and the NFA frameworks), and different function approximation
‘ The bold numbers in the table indicate the best results.
‘ The bold numbers in the table indicate the best results.
FIGURE 3.1: Block diagram of the proposed speaker height estimation approach in training and testing phases.
FIGURE 3.1: Block diagram of the proposed speaker height estimation approach in training and testing phases.
FIGURE 3.2: The height histogram of telephone speech utterances for the NIST 2008 and NIST 2010 databases.
FIGURE 3.2: The height histogram of telephone speech utterances for the NIST 2008 and NIST 2010 databases.
FIGURE 3.3: The scatter plot of height estimation for (a): male speakers, (b): female speakers, and (c): when the male and female data were pooled together.
FIGURE 3.3: The scatter plot of height estimation for (a): male speakers, (b): female speakers, and (c): when the male and female data were pooled together.
FIGURE 4.1: Block diagram of the proposed speaker weight estimation approach in training and testing phases.
FIGURE 4.1: Block diagram of the proposed speaker weight estimation approach in training and testing phases.
FIGURE 4.2: The weight histogram of telephone speech utterances of training and testing datasets for male and female speaker.
FIGURE 4.2: The weight histogram of telephone speech utterances of training and testing datasets for male and female speaker.
* The bold numbers in the table indicate the best results. TABLE 4.1: Results of speaker weight estimation using MLPs and LSSVR. CC is the Pearson correlation coefficient between actual and estimated weight, and MAE is the mean-absolute-error between actual and estimated weight, in ke.
* The bold numbers in the table indicate the best results. TABLE 4.1: Results of speaker weight estimation using MLPs and LSSVR. CC is the Pearson correlation coefficient between actual and estimated weight, and MAE is the mean-absolute-error between actual and estimated weight, in ke.
FIGURE 5.2: Block-diagram of the proposed smoker detection approach in training and testing phases. 5. AUTOMATIC SMOKER DETECTION FROM SPONTANEOUS TELEPHONE SPEECH
FIGURE 5.2: Block-diagram of the proposed smoker detection approach in training and testing phases. 5. AUTOMATIC SMOKER DETECTION FROM SPONTANEOUS TELEPHONE SPEECH
FIGURE 5.3: The smoking habit histogram of telephone speech utterances for training, development and testing datasets for (a): male speakers and (b): female speakers.
FIGURE 5.3: The smoking habit histogram of telephone speech utterances for training, development and testing datasets for (a): male speakers and (b): female speakers.
TABLE 5.1: The Cimin of applying different classifiers (Multilayer Perceptrons (MLP), Logistic Regression (LR), Von-Mises-Fisher Scoring (VMF), Gaussian Scoring (GS) and Naive Bayesian Classifier (NBC)) over the i-vector and the NFA frameworks
TABLE 5.1: The Cimin of applying different classifiers (Multilayer Perceptrons (MLP), Logistic Regression (LR), Von-Mises-Fisher Scoring (VMF), Gaussian Scoring (GS) and Naive Bayesian Classifier (NBC)) over the i-vector and the NFA frameworks
9.4 Results and Discussion This section presents the results of the proposed smoking habit detection approach. The acoustic feature consists of 20 Mel-Frequency Cepstrum Coefficients (MFCCs) including energy appended with their first and second order derivatives, forming a 60 dimensional acoustic feature vector. MFCCs are obtained using cosine transform of the real logarithm of the short-term energy spectrum represented on a mel-frequency scale [80]. This type of feature is very common in the state-of-the-art i-vector-based speaker recognition systems. To have more reliable features, Wiener filtering, speech activity detection [68] and feature warping [80] have been considered in front-end processing. This section presents the results of the proposed smoking habit detection approach.
9.4 Results and Discussion This section presents the results of the proposed smoking habit detection approach. The acoustic feature consists of 20 Mel-Frequency Cepstrum Coefficients (MFCCs) including energy appended with their first and second order derivatives, forming a 60 dimensional acoustic feature vector. MFCCs are obtained using cosine transform of the real logarithm of the short-term energy spectrum represented on a mel-frequency scale [80]. This type of feature is very common in the state-of-the-art i-vector-based speaker recognition systems. To have more reliable features, Wiener filtering, speech activity detection [68] and feature warping [80] have been considered in front-end processing. This section presents the results of the proposed smoking habit detection approach.
FIGURE 5.4: Block-diagram of the proposed smoker detection approach for score- level fusion of the i-vector-based recognizer (model-1) and the NFA-based recognizer (model-2). (U.M. stands for utterance modeling)
FIGURE 5.4: Block-diagram of the proposed smoker detection approach for score- level fusion of the i-vector-based recognizer (model-1) and the NFA-based recognizer (model-2). (U.M. stands for utterance modeling)
FiGuRE 5.5: The ROC curves of the proposed method for (a): male speakers, and (b): female speakers.
FiGuRE 5.5: The ROC curves of the proposed method for (a): male speakers, and (b): female speakers.
FIGURE 6.1: The scatter plots of (a): age and height, (b): age and weight, and (c): height and weight of speakers in NIST 2008 and 2010 SRE databases for both genders.
FIGURE 6.1: The scatter plots of (a): age and height, (b): age and weight, and (c): height and weight of speakers in NIST 2008 and 2010 SRE databases for both genders.
FIGURE 6.2: Block diagram of the proposed multitask speaker profiling approach for speaker age estimation and smoking habit detection, in training, development and testing phases. U.M. stands for utterance modeling. yf and y?. represent the training labels corresponding to age and smoking habit, respectively, and g4 and 9° represent the estimated age and smoking habit, respectively, after applying a test sample 2+.+.
FIGURE 6.2: Block diagram of the proposed multitask speaker profiling approach for speaker age estimation and smoking habit detection, in training, development and testing phases. U.M. stands for utterance modeling. yf and y?. represent the training labels corresponding to age and smoking habit, respectively, and g4 and 9° represent the estimated age and smoking habit, respectively, after applying a test sample 2+.+.
FIGURE 6.3: Block diagram of the proposed multitask speaker profiling approach for speaker age, height and weight estimation, in training and testing phases. U.M. stands for utterance modeling. yf, yf! and y}{" represent the training labels corresponding to age, height and weight, respectively, and g4, g” and g” represent the estimated age, height and weight, respectively, after applying a test sample 7;.;. speaker age, height and weight estimation, in training and testing phases. U.M. stands
FIGURE 6.3: Block diagram of the proposed multitask speaker profiling approach for speaker age, height and weight estimation, in training and testing phases. U.M. stands for utterance modeling. yf, yf! and y}{" represent the training labels corresponding to age, height and weight, respectively, and g4, g” and g” represent the estimated age, height and weight, respectively, after applying a test sample 7;.;. speaker age, height and weight estimation, in training and testing phases. U.M. stands
‘ The bold numbers in the table indicate the improved results. TABLE 6.1: The comparison between single-task and multitask speaker profiling for speaker height and age estimation. C'C is the Pearson correlation coefficient between actual and estimated height /age.
‘ The bold numbers in the table indicate the improved results. TABLE 6.1: The comparison between single-task and multitask speaker profiling for speaker height and age estimation. C'C is the Pearson correlation coefficient between actual and estimated height /age.
The bold numbers in the table indicate the improved results. TABLE 6.2: The comparison between single-task and multitask speaker profiling for speaker weight and age estimation. CC is the Pearson correlation coefficient between actual and estimated weight /age
The bold numbers in the table indicate the improved results. TABLE 6.2: The comparison between single-task and multitask speaker profiling for speaker weight and age estimation. CC is the Pearson correlation coefficient between actual and estimated weight /age
TABLE 6.3: The comparison between single-task and multitask speaker profilins for speaker height and weight estimation. C'C is the Pearson correlation coefficien between actual and estimated height /weight TABLE 6.4: The comparison between single-task and multitask speaker profiling for speaker height, weight and age estimation. C'C is the Pearson correlation coefficient between actual and estimated height /weight /age
TABLE 6.3: The comparison between single-task and multitask speaker profilins for speaker height and weight estimation. C'C is the Pearson correlation coefficien between actual and estimated height /weight TABLE 6.4: The comparison between single-task and multitask speaker profiling for speaker height, weight and age estimation. C'C is the Pearson correlation coefficient between actual and estimated height /weight /age
Pearson correlation coefficient between actual and estimated age
Pearson correlation coefficient between actual and estimated age
FIGURE 6.4: The ROC curves of the proposed MTL smoker detection after the score-level fusion and when age information is considered, for (a): male speakers, and (b): female speakers. * The bold numbers in the table indicate the improved results.
FIGURE 6.4: The ROC curves of the proposed MTL smoker detection after the score-level fusion and when age information is considered, for (a): male speakers, and (b): female speakers. * The bold numbers in the table indicate the improved results.
TABLE 7.2: The relative improvements (R.I.) in Cipmin and AUC of the proposed smoking habit detection after score-level fusion compared with the i-vector framework.
TABLE 7.2: The relative improvements (R.I.) in Cipmin and AUC of the proposed smoking habit detection after score-level fusion compared with the i-vector framework.
the baseline [9] (provided on the same databases) for males and females by 8.6% and 22.2%, respectively, which reflects the effectiveness of the proposed method in automatic speaker age estimation. the baseline |9| (provided on the same databases) for males and females by 8.6%
the baseline [9] (provided on the same databases) for males and females by 8.6% and 22.2%, respectively, which reflects the effectiveness of the proposed method in automatic speaker age estimation. the baseline |9| (provided on the same databases) for males and females by 8.6%
TABLE 7.3: The relative improvements in CC for age, height and weight estimations obtained in a multitask age, height and weight estimation, compared with the baselines. TABLE 7.4: The relative improvements (R.I.) in Cyipmin and AUC for smoking habit detection and the R.I. in CC for age estimation obtained in a multitask smoking habit and age estimation compared with baselines. TABLE 7.4: The relative improvements (R.I.) in Clip min and AUC for smoking habit
TABLE 7.3: The relative improvements in CC for age, height and weight estimations obtained in a multitask age, height and weight estimation, compared with the baselines. TABLE 7.4: The relative improvements (R.I.) in Cyipmin and AUC for smoking habit detection and the R.I. in CC for age estimation obtained in a multitask smoking habit and age estimation compared with baselines. TABLE 7.4: The relative improvements (R.I.) in Clip min and AUC for smoking habit
profiling systems was investigated. In the proposed method, each utterance of the NIST 2008 and 2010 SRE databases was modeled using the i-vector and the NFA frameworks. In this study, due to the task relatedness, a MTL for speaker age, height and weight estimation, and a MTL for speaker age and smoking habit estimation were performed in separate experiments.
profiling systems was investigated. In the proposed method, each utterance of the NIST 2008 and 2010 SRE databases was modeled using the i-vector and the NFA frameworks. In this study, due to the task relatedness, a MTL for speaker age, height and weight estimation, and a MTL for speaker age and smoking habit estimation were performed in separate experiments.
FicurE A.2: The structure of a multilayer perceptron (MLP) The activation functions commonly used in feedforward neural networks are logistic, hyperbolic tangent and linear functions. The logistic function takes the following form, which its output lies between 0 and 1:
FicurE A.2: The structure of a multilayer perceptron (MLP) The activation functions commonly used in feedforward neural networks are logistic, hyperbolic tangent and linear functions. The logistic function takes the following form, which its output lies between 0 and 1:
FIGURE A.3: The structure of a single hidden layer feedforward neural network with error back propagation. The solid lines represent the forward paths and the dotted lines indicate the error back-propagation paths.
FIGURE A.3: The structure of a single hidden layer feedforward neural network with error back propagation. The solid lines represent the forward paths and the dotted lines indicate the error back-propagation paths.
In the binary classification problems, we intend to model the probability of a T certain label given its featurs. That is, P(y|x;) = f(w* x; + wo), where x; is the feature vector corresponding to the i*” sample. Vector w and constant wo are the model parameters, which are found through the maximum likelihood estimation (MLI next E). The MLI section. E in the logistic regression model for binary cases is described in the The output of the logistic function, as shown in Figure (C.1), takes a value between zero and one.
In the binary classification problems, we intend to model the probability of a T certain label given its featurs. That is, P(y|x;) = f(w* x; + wo), where x; is the feature vector corresponding to the i*” sample. Vector w and constant wo are the model parameters, which are found through the maximum likelihood estimation (MLI next E). The MLI section. E in the logistic regression model for binary cases is described in the The output of the logistic function, as shown in Figure (C.1), takes a value between zero and one.
At this moment, since the a is a positive and smaller than one, the B matrix is positive semidefinite,
At this moment, since the a is a positive and smaller than one, the B matrix is positive semidefinite,
Related topics:
Speaker RecognitionSpeech ProcessingSpeaker CharacterizationI-Vector
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved