Acoustic feature selection and classification of emotions in speech using a 3D continuous emotion model
Biomedical Signal Processing and ControlPublished 8 June 2011
Humberto Pérez-Espinosa, Carlos A. Reyes-García, Luis Villaseñor-Pineda
Citations37
SJR quartileQ1
SJR score1.23
SNIP1.65
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
The feasibility of applying the continuous emotion models approach to annotation of emotional speech is demonstrated and ways to take advantage of this kind of annotation to improve the automatic classification of basic emotions are explored.
Abstract
In this paper we report the results obtained from experiments with a database of emotional speech in English in order to find the most important acoustic features to estimate Emotion Primitives which determine the emotional content on speech. We are interested in exploiting the potential benefits of continuous emotion models, so in this paper we demonstrate the feasibility of applying this approach to annotation of emotional speech and we explore ways to take advantage of this kind of annotation to improve the automatic classification of basic emotions.
Keywords
PsychologyComputer ScienceHealth Professions
Language Resources and EvaluationIEMOCAP: interactive emotional dyadic motion capture database
3,632 Citations2008Carlos Busso, Murtaza Bulut +7 more
A new corpus named the “interactive emotional dyadic motion capture database” (IEMOCAP), collected by the Speech Analysis and Interpretation Laboratory at the University of Southern California (USC), which provides detailed information about their facial expressions and hand movements during scripted and spontaneous spoken communication scenarios.
Pattern Recognition LettersFloating search methods in feature selection
3,070 Citations1994Pavel Pudil, Jana Novovičová +1 more
A database of German emotional speech
2,200 Citations2005Felix Burkhardt, Astrid Paeschke +3 more
A database of emotional speech that was evaluated in a perception test regarding the recognisability of emotions and their naturalness and can be accessed by the public via the internet.
Giornale italiano di dermatologia e venereologia : organo ufficiale, Società italiana di dermatologia e sifilografiaData Mining: Practical Machine Learning Tools and Techniques, Second Edition (Morgan Kaufmann Series in Data Management Systems)
817 Citations2005Ian H. Witten, Eibe Frank
OpenEAR — Introducing the munich open-source emotion and affect recognition toolkit
410 Citations2009Florian Eyben, Martin Wöllmer +1 more
A novel open-source affect and emotion recognition engine, which integrates all necessary components in one highly efficient software package, and which can be used for batch processing of databases.
The Vera am Mittag German audio-visual emotional speech database
406 Citations2008Michael Grimm, Kristian Kroschel +1 more
This contribution presents a recently collected database of spontaneous emotional speech in German which is being made available to the research community and provides emotion labels for a great part of the data.
Cognition Technology & WorkEmotion recognition from physiological signals using wireless sensors for presence technologies
365 Citations2004Fatma Nasoz, Kaye Alvarez +2 more
A new approach to enhance presence technologies that will adapt to user affect dynamically in the current context, thus providing enhanced social presence and introducing the prototype multimodal affective user interface.
Large-scale attribute selection using wrappers
266 Citations2009Martin Gütlein, Eibe Frank +2 more
This work investigates Linear Forward Selection, a technique to reduce the number of attributes expansions in each forward selection step and shows that this approach is faster, finds smaller subsets and can even increase the accuracy compared to standard forward selection.
World Congress on Medical Physics and Biomedical Engineering, September 7 - 12, 2009, Munich, Germany4th Kuala Lumpur International Conference on Biomedical Engineering 2008
146 Citations2008Magjarevic, R, Wan Abas, Wan Abu Bakar +6 more
Computer Speech & LanguageWhodunnit – Searching for the most important feature types signalling emotion-related user states in speech
146 Citations2010Anton Batliner, Stefan Steidl +9 more
A set of acoustic and linguistic features that characterise emotional/emotion-related user states - confined to the one database processed: four classes in a German corpus of children interacting with a pet robot are described and interpreted.
Journal on Multimodal User InterfacesOn-line emotion recognition in a 3-D activation-valence-time continuum using acoustic and linguistic cues
140 Citations2009Florian Eyben, Martin Wöllmer +4 more
This work presents a novel approach to on-line emotion recognition from speech using Long Short-Term Memory Recurrent Neural Networks, which recognition is performed on low-level signal frames, similar to those used for speech recognition.
Decision level combination of multiple modalities for recognition and analysis of emotional expression
93 Citations2010Angeliki Metallinou, Sungbok Lee +1 more
This work model face, voice and head movement cues for emotion recognition and fuse classifiers using a Bayesian framework and suggests a positive correlation between the number of classifiers that performed well and the perceptual salience of the expressed emotion.
Lecture notes in computer scienceFrame vs. Turn-Level: Emotion Recognition from Speech Considering Static and Dynamic Processing
86 Citations2007Bogdan Vlasenko, Björn W. Schuller +2 more
This work re-investigate dynamic modeling directly on the frame-level in speech-based emotion recognition and integrates this frame- level information within a state-of-the-art large-feature-space emotion recognition engine.
Cepstral and long-term features for emotion recognition
57 Citations2009Pierre Dumouchel, Najim Dehak +3 more
Systems that were developed for the Open Performance Sub-Challenge of the INTERSPEECH 2009 Emotion Challenge are described, which participate in both two-class and fiveclass emotion detection.
Acta Otorrinolaringológica EspañolaEvaluación perceptual de la disfonía: correlación con los parámetros acústicos y fiabilidad
52 Citations2004Faustino Núñez-Batalla, P. Corte Santos +3 more
The results of this study show that GRBAS test-retest reliability and intrerrater agreement is high and the severity of dysphonia was assesed better by breath related parameters and low fundamental frequencies.
Data-driven clustering in emotional space for affect recognition using discriminatively trained LSTM networks
50 Citations2009Martin Wöllmer, Florian Eyben +3 more
It is shown that a discriminatively trained Long Short-Term Memory (LSTM) recurrent neural net that explicitly learns clusters in emotional space and additionally models context information outperforms both, Support Vector Machines and a Regression-L STM net.
International Journal of Speech TechnologyVocal emotion recognition in five native languages of Assam using new wavelet features
45 Citations2009Aditya Bihar Kandali, Aurobinda Routray +1 more
An emotional utterance database is created with 140 acted utterances per speaker consisting of short sentences of six full-blown basic emotions and neutral of five native languages of Assam to help in developing a generalized voice emotion recognition system, which will increase the efficiency of human-machine interaction systems.
Lecture notes in computer scienceReal-Life Emotion Representation and Detection in Call Centers Data
42 Citations2005Laurence Vidrascu, Laurence Devillers
This paper addresses some of the challenges faced when studying real-life non-basic emotions in spoken dialogs from call-center services and uses a soft emotion vector to represent emotion mixtures.
EURASIP Journal on Advances in Signal ProcessingOn the Use of the Correlation between Acoustic Descriptors for the Normal/Pathological Voices Discrimination
39 Citations2009Thomas Dubuisson, Thierry Dutoit +2 more
Based on the normal and pathological samples included the MEEI database, it has been found that using two features and their correlation leads to correct classification rates of 94.7% for pathological voices and 89.5% for normal ones.
Proceedings of the ... IEEE International Conference on Acoustics, Speech, and Signal ProcessingCascaded emotion classification via psychological emotion dimensions using a large set of voice quality parameters
39 Citations2008Marko Lugger, Bin Yang
This paper improves the speaker independent emotion classification of a well known German database consisting of 6 basic emotions by adding a large set of voice quality parameters to the standard prosodic features, leading to a 3-stage cascaded classification motivated by the psychological model of emotion dimensions.
Improving automatic emotion recognition from speech signals
34 Citations2009Elif Bozkurt, Engin Erzin +2 more
A speech signal driven emotion recognition system trained and tested with the INTERSPEECH 2009 Emotion Challenge corpus, which includes spontaneous and emotionally rich recordings, and investigates prosody related, spectral and HMM-based features for the evaluation of emotion recognition with Gaussian mixture model (GMM) based classifiers.
AN INCREMENTAL ANALYSIS OF DIFFERENT FEATURE GROUPS IN SPEAKER INDEPENDENT EMOTION RECOGNITION
33 Citations2007Marko Lugger, Bin Yang
This paper investigates the classification of different emotional states using speech features from different feature groups and uses a Bayesian classifier and a speaker independent cross validation to improve the classification performance of speaker independent emotion recognition by combining conventional prosodic features with additional segmental features.
Features selection for primitives estimation on emotional speech
31 Citations2010Humberto Pérez-Espinosa, Carlos Alberto Reyes García +1 more
This paper examines the importance of different groups of speech acoustic features in the estimation of emotional primitives which define a three-dimensional continuous model of emotions.
Logopedics Phoniatrics VocologyNoh voice quality
27 Citations2009Osamu Fujimura, Kiyoshi Honda +4 more
Using a signal analysis-synthesis algorithm, TANDEM-STRAIGHT, to represent multiple candidates for pitch perception, signals deviating from the harmonic structure have been successfully displayed, corresponding to auditory impressions of pitch movements, even when narrow-band spectrograms failed to show the perceived events.
Proposal of acoustic measures for automatic detection of vocal fry
22 Citations2005Carlos Toshinori Ishi, Hiroshi Ishiguro +1 more
A set of acoustic measures for automatically detecting vocal fry segments in speech utterances based on power, periodicity and similarity properties of vocal fry signals are proposed.
The auditory processing and recognition of speech
10 Citations1989William Byrne, John Robinson +1 more
Studies of the robustness of a speech representation using a biophysical model of the cochlea, experimental results on the representation of speech and complex sounds in the mammalian auditory cortex, and descriptions of computational sequential processing networks capable of recognizing sequences of phonemes are described.
Estrategias para la mejora de la naturalidad y la incorporación de variedad emocional a la conversión texto a voz en castellano
9 Citations2011Montero Martínez, Juan Manuel
Philips ResearchA Wearable Emg Monitoring System for Emotions Assessment
8 Citations2007Cecilia Vera‐Muñoz, Laura Pastor‐Sanz +4 more
An innovative, intelligent, multi-sensor and wearable system for the assessment of the emotional state of humans under special conditions (i.e. neurological diseases, stress, etc.) and can be applied in diverse areas.
Lecture notes in computer scienceStatistical Feature Selection for Mandarin Speech Emotion Recognition
5 Citations2005Bo Xie, Ling Chen +2 more
The features selected by SDA are better than the results of other feature selection methods in a LDA based classifier and SVM and the best performance is achieved when the feature number is in the range of 9 to 12.