Powered by OpenAIRE graph
Found an issue? Give us feedback
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/ UPCommons. Portal de...arrow_drop_down
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
Recolector de Ciencia Abierta, RECOLECTA
Bachelor thesis . 2017
License: CC BY NC ND
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
Recolector de Ciencia Abierta, RECOLECTA
Bachelor thesis . 2017
License: CC BY NC ND
versions View all 4 versions
addClaim

Multimodal expression analysis

Análisis multimodal de expresiones
Authors: Aguilera Martínez, Carlos;

Multimodal expression analysis

Abstract

El uso del reconocimiento de emociones ha ido en aumento en los últimos años, desde las redes sociales y los videojuegos a la investigación sobre la experiencia del cliente. Este hecho ha creado un interés en la comunidad para su investigación y desarrollo. Este trabajo se ha centrado en el estudio del uso de dos modalidades, habla y video, para entrenar una CNN para determinar una emoción de una secuencia de vídeo. Se ha tomado un enfoque de deep learning para las características de video y clasificadores clásicos para audio. El sistema ha sido entrenado principalmente con la base de datos AFEW lo que hace que adopte un enfoque adaptado al entorno real en lugar de un entorno de laboratorio. Esto hace que la detección de emociones sea más difícil debido a la cantidad de ruido en ambas modalidades. Se presentan tres sistemas con fusión multimodal a nivel de decisión, y dos sistemas a nivel de fusión de características.

L'ús del reconeixement d'emocions ha anat en augment en els últims anys, des de les xarxes socials i els videojocs a la investigació sobre l'experiència del client. Aquest fet ha creat un interès en la comunitat per a la seva investigació i desenvolupament. Aquest treball s'ha centrat en l'estudi de l'ús de dues modalitats, parla i vídeo, per entrenar una CNN per determinar una emoció d'una seqüència de vídeo. S'ha pres un enfocament de deep learning per les característiques de i classificadors clàssics per àudio. El sistema ha estat entrenat principalment amb la base de dades AFEW el que fa que adopti un enfocament adaptat a l'entorn real en lloc d'un entorn de laboratori. Això fa que la detecció d'emocions sigui més difícil a causa de la quantitat de soroll en els dos modes. Es presenten tres sistemes amb fusió multimodal a nivell de decisió, i dos sistemes a nivell de fusió de característiques.

The use for emotion recognition has been on the rise on the last few years, from social media and video games to the research on customer's experience. This fact has created an interest on the community for its research and development. This work has focused on the study of the use of two modalities, speech and video, to train a CNN to determine an emotion out of a video sequence. A deep learning approach has been used for video features and classical classifiers for audio. The system has been trained mainly with the AFEW database which makes it take an approach adapted to the real environment instead of a lab environment. This makes the emotion detection more challenging due to the amount of noise in both the modalities. Three systems are presented with multimodal fusion at decision level, and two systems at feature fusion level.

Country
Spain
Keywords

Neural networks (Computer science), Machine Learning, Redes convolucionales, Aprenentatge automàtic, Xarxes neuronals (Informàtica), Deep learning, Reconociemiento de emociones, Emotion recognition, :Enginyeria de la telecomunicació [Àrees temàtiques de la UPC], Àrees temàtiques de la UPC::Enginyeria de la telecomunicació, CNN

  • BIP!
    Impact byBIP!
    selected citations
    These citations are derived from selected sources.
    This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    0
    popularity
    This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
    Average
    influence
    This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    Average
    impulse
    This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
    Average
    OpenAIRE UsageCounts
    Usage byUsageCounts
    visibility views 28
    download downloads 24
  • 28
    views
    24
    downloads
    Powered byOpenAIRE UsageCounts
Powered by OpenAIRE graph
Found an issue? Give us feedback
visibility
download
selected citations
These citations are derived from selected sources.
This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Citations provided by BIP!
popularity
This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
BIP!Popularity provided by BIP!
influence
This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Influence provided by BIP!
impulse
This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
BIP!Impulse provided by BIP!
views
OpenAIRE UsageCountsViews provided by UsageCounts
downloads
OpenAIRE UsageCountsDownloads provided by UsageCounts
0
Average
Average
Average
28
24
Green