skip to main content
Idioma:

PERFORMANCE EVALUATION OF VARIOUS EMOTION CLASSIFICATION APPROACHES FROM PHYSIOLOGICAL SIGNALS

Geroge Patton ; Zhang, Lin

figshare 2018

Texto completo disponível

Citações Citado por
  • Título:
    PERFORMANCE EVALUATION OF VARIOUS EMOTION CLASSIFICATION APPROACHES FROM PHYSIOLOGICAL SIGNALS
  • Autor: Geroge Patton ; Zhang, Lin
  • Assuntos: Artificial Intelligence and Image Processing ; Artificial Intelligence and Image Processing not elsewhere classified ; Artificial Life ; FOS: Computer and information sciences
  • Descrição: This paper aims at evaluating the performance of various emotion classification approaches from psychophysiological signals. The goal is to identify the combinations of approaches that are most relevant for assessing human affective states. A classification analysis of various combinations of feature selection techniques, classification algorithms and evaluation methods is presented. The emotion recognition is conducted based on four physiological signals: two electromyograms, skin conductivity and respiration sensors. Affective states are classified into three different emotion classes: 2-categoryclass (Arousal), 3-category-class (Valence) and 5-category-class (Valence/Arousal). The performance of the various combinations of approaches is evaluated by comparing the resulting recognition rates. For all the category-classes, the best results are obtained when considering skin conductivity combined with the respiration signals. Highest rates when fusing all physiological channels resulted when applying the SFS feature selection, the LDA classifier and the normal split evaluation approach, showing a robust combination of approaches leading to good performance.
  • Editor: figshare
  • Data de criação/publicação: 2018
  • Idioma: Inglês
Links
Voltar para lista de resultados

  • Realização: ABCD-USP USP   Logos de Redes Sociais: Freepik

Buscando em bases de dados remotas. Favor aguardar.