Efficient speech emotion recognition using modified feature extraction

Langari, Shadi; Marvi, Hossein; Zahedi, Morteza

doi:10.1016/j.imu.2020.100424

Cited by 25 publications

(7 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In essence, data classification investigates the relations between feature variables (i.e., inputs) and output variables. Classification methods have been used in a broad range of applications such as customer target marketing [1,2], medical disease diagnosis [3][4][5], speech and handwriting recognition [6][7][8][9], multimedia data analysis [10,11], biological data analysis [12], document categorization and filtering [13,14], and social network analysis [15][16][17]. Classification algorithms typically contain two steps, the learning step and the testing step.…”

Section: Classification Methodsmentioning

confidence: 99%

Ontology-Based Feature Selection: A Survey

2021

View full text Add to dashboard Cite

The Semantic Web emerged as an extension to the traditional Web, adding meaning (semantics) to a distributed Web of structured and linked information. At its core, the concept of ontology provides the means to semantically describe and structure information, and expose it to software and human agents in a machine and human-readable form. For software agents to be realized, it is crucial to develop powerful artificial intelligence and machine-learning techniques, able to extract knowledge from information sources, and represent it in the underlying ontology. This survey aims to provide insight into key aspects of ontology-based knowledge extraction from various sources such as text, databases, and human expertise, realized in the realm of feature selection. First, common classification and feature selection algorithms are presented. Then, selected approaches, which utilize ontologies to represent features and perform feature selection and classification, are described. The selective and representative approaches span diverse application domains, such as document classification, opinion mining, manufacturing, recommendation systems, urban management, information security systems, and demonstrate the feasibility and applicability of such methods. This survey, in addition to the criteria-based presentation of related works, contributes a number of open issues and challenges related to this still active research topic.

show abstract

Section: Classification Methodsmentioning

confidence: 99%

Ontology-Based Feature Selection: A Survey

2021

View full text Add to dashboard Cite

show abstract

“…The framing of the speech signal is realized by the process of weighting with a window function. There are usually two methods of framing, continuous segmentation and overlapping segmentation [28]. However, to ensure a smooth transition between adjacent frames, overlapping segmentation is usually adopted.…”

Section: Preprocessingmentioning

confidence: 99%

“…After data standardization, all indicators are in the same order of magnitude, and they are suitable for comprehensive evaluation. After feature normalization, the contour of the loss function will be a partial circle, the gradient descent process will be flatter [28], and the convergence will be faster; therefore, the performance will be improved. The StandardScaler [34] is used to normalize the data so that the new dataset has a mean and standard deviation of zero.…”

Section: Feature Extractionmentioning

confidence: 99%

Research on Speech Emotion Recognition Method Based A-CapsNet

Huang

Zhang

2022

Applied Sciences

View full text Add to dashboard Cite

Speech emotion recognition is a crucial work direction in speech recognition. To increase the performance of speech emotion detection, researchers have worked relentlessly to improve data augmentation, feature extraction, and pattern formation. To address the concerns of limited speech data resources and model training overfitting, A-CapsNet, a neural network model based on data augmentation methodologies, is proposed in this research. In order to solve the issue of data scarcity and achieve the goal of data augmentation, the noise from the Noisex-92 database is first combined with four different data division methods (emotion-independent random-division, emotion-dependent random-division, emotion-independent cross-validation and emotion-dependent cross-validation methods, abbreviated as EIRD, EDRD, EICV and EDCV, respectively). The database EMODB is then used to analyze and compare the performance of the model proposed in this paper under different signal-to-noise ratios, and the results show that the proposed model and data augmentation are effective.

show abstract

“…Different techniques in SER classification have been constantly developed and improved over the years. Some have extracted novel types of features like adaptive time-frequency features [16] based on the fractional Fourier transformation and frequency modulation features [17] based on the amplitude modulation-frequency modulation model. In contrast to designing new kinds of features,Özseven [18] instead proposes a novel featureselection method.…”

Section: Relevant Theoretical Bases and Literaturementioning

confidence: 99%

The Investigation of Different Loss Functions with Capsule Networks for Speech Emotion Recognition

Liu

2021

Scientific Programming

View full text Add to dashboard Cite

Speech emotion recognition (SER) is an important research topic. Image features like spectrograms are one of the common ways of extracting information from speech. In the area of image recognition, a relatively novel type of network called capsule networks has shown good and promising results. This study aims to use capsule networks to encode spatial information from spectrograms and analyse its performance when paired with different loss functions. Experiments comparing the capsule network with models from previous works show that the capsule network performs better than them.

show abstract

Efficient speech emotion recognition using modified feature extraction

Cited by 25 publications

References 25 publications

Ontology-Based Feature Selection: A Survey

Ontology-Based Feature Selection: A Survey

Research on Speech Emotion Recognition Method Based A-CapsNet

The Investigation of Different Loss Functions with Capsule Networks for Speech Emotion Recognition

Contact Info

Product

Resources

About