Fuzzy emotion recognition in natural speech dialogue

Austermann, Anja; Esau, Natascha; Kleinjohann, Lisa; Kleinjohann, Bernd

doi:10.1109/roman.2005.1513798

Cited by 23 publications

(10 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The cross-corpora evaluation in Shami and Verhels [15] demonstrated the drop in classification performance observed when training on one emotional corpus and testing on another. Other studies have shown similar results [5], [16]- [18]. Several approaches have been proposed to solve this problem.…”

Section: Related Worksupporting

confidence: 56%

Domain Adversarial for Acoustic Emotion Recognition

Abdelwahab

Busso

2018

IEEE/ACM Trans. Audio Speech Lang. Process.

150

View full text Add to dashboard Cite

The performance of speech emotion recognition is affected by the differences in data distributions between train (source domain) and test (target domain) sets used to build and evaluate the models. This is a common problem, as multiple studies have shown that the performance of emotional classifiers drop when they are exposed to data that does not match the distribution used to build the emotion classifiers. The difference in data distributions becomes very clear when the training and testing data come from different domains, causing a large performance gap between validation and testing performance. Due to the high cost of annotating new data and the abundance of unlabeled data, it is crucial to extract as much useful information as possible from the available unlabeled data. This study looks into the use of adversarial multitask training to extract a common representation between train and test domains. The primary task is to predict emotional attribute-based descriptors for arousal, valence, or dominance. The secondary task is to learn a common representation where the train and test domains cannot be distinguished. By using a gradient reversal layer, the gradients coming from the domain classifier are used to bring the source and target domain representations closer. We show that exploiting unlabeled data consistently leads to better emotion recognition performance across all emotional dimensions. We visualize the effect of adversarial training on the feature representation across the proposed deep learning architecture. The analysis shows that the data representations for the train and test domains converge as the data is passed to deeper layers of the network. We also evaluate the difference in performance when we use a shallow neural network versus a deep neural network (DNN) and the effect of the number of shared layers used by the task and domain classifiers.Index Terms-Speech emotion recognition, adversarial training, unlabeled adaptation of acoustic emotional models.

show abstract

Section: Related Worksupporting

confidence: 56%

Domain Adversarial for Acoustic Emotion Recognition

Abdelwahab

Busso

2018

IEEE/ACM Trans. Audio Speech Lang. Process.

150

View full text Add to dashboard Cite

show abstract

“…For example, Razak [11] used 22 emotion features to recognize emotions, and the average recognition rate was 68.59%. Austermann [13] has done research on robot emotion recognition in the natural speech dialogue, membership functions were used to select the most significant features from a set of twenty analyzed features. Austermann [13] has done research on robot emotion recognition in the natural speech dialogue, membership functions were used to select the most significant features from a set of twenty analyzed features.…”

Section: Related Work Comparisonmentioning

confidence: 99%

“…Austermann [13] has done research on robot emotion recognition in the natural speech dialogue, membership functions were used to select the most significant features from a set of twenty analyzed features. Moreover, in speaker-independent system, the average recognition rate for six emotion states we got is higher than that for four emotion states in paper [13]. In this paper, with the method we proposed, we get a high average recognition rate of 78.6% for six emotion states in speaker-independent system.…”

Section: Related Work Comparisonmentioning

confidence: 99%

Speech Emotion Feature Selection Method Based on Contribution Analysis Algorithm of Neural Network

Wang¹,

Mao²,

Zhan³

2008

AIP Conference Proceedings

View full text Add to dashboard Cite

Articles you may be interested inNSDann2BS, a neutron spectrum unfolding code based on neural networks technology and two bonner spheres AIP Conf. Proc. 1544, 122 (2013); 10.1063/1.4813469Evaluating the performance of two neutron spectrum unfolding codes based on iterative procedures and artificial neural networks AIP Conf.From neural to quantum associative networks: A new quantum "algorithm" AIP Conf.Abstract. There are many emotion features. If all these features are employed to recognize emotions, redundant features may be existed. Furthermore, recognition result is unsatisfying and the cost of feature extraction is high. In this paper, a method to select speech emotion features based on contribution analysis algorithm of NN is presented. The emotion features are selected by using contribution analysis algorithm of NN from the 95 extracted features. Cluster analysis is applied to analyze the effectiveness for the features selected, and the time of feature extraction is evaluated. Finally, 24 emotion features selected are used to recognize six speech emotions. The experiments show that this method can improve the recognition rate and the time of feature extraction In this paper, 23.2 ms (256 points) frames of speech signal are analyzed every lOms overlap with Hamming window. In order to make full use of emotion information of speech signals, and in terms of characteristic of

show abstract

“…Otros propósitos del uso de señales de voz es la formulación e implementación de sistemas para la distinción del habla [7]- [9], reconocimiento del habla [10], reconocimiento de lenguaje [11], reconocimiento de emociones basados en habla y en género [12], [13], diagnóstico de enfermedades patológicas como la disfonía y la laringitis [14], diagnóstico de nódulos, edemas y parálisis unilateral de las cuerdas vocales [15], reconocimiento de disartria en sistemas de reconocimiento automático de habla [16] y la detección temprana de Parkinson mediante voz [17].…”

Section: Introductionunclassified

Gender classification based on voice signals using fuzzy models and optimization algorithms

Martinez

Espitia

2019

ITECKNE

View full text Add to dashboard Cite

This paper describes a gender classification scheme based on voice signals in which 16 different fuzzy models are proposed and optimized using four bio-inspired optimization algorithms and the quasi-Newton method. The classification scheme considers four data sets and five different voice features to define the input values of an algorithm in the optimization process. The inputs of each fuzzy model define the mean and variance of their Gaussian membership functions, and their fitness is evaluated by the input values of the algorithm and mean squared error as objective function to be minimized. A comparative analysis between models, algorithms and data sets is made to obtain conclusions according to the results of each optimized model.

show abstract

Fuzzy emotion recognition in natural speech dialogue

Cited by 23 publications

References 8 publications

Domain Adversarial for Acoustic Emotion Recognition

Domain Adversarial for Acoustic Emotion Recognition

Speech Emotion Feature Selection Method Based on Contribution Analysis Algorithm of Neural Network

Gender classification based on voice signals using fuzzy models and optimization algorithms

Contact Info

Product

Resources

About