Recognizing emotions expressed by body pose: A biologically inspired neural model

被引：126

作者：

Schindler, Konrad ^{[1
]}

Van Gool, Luc ^{[1
,2
]}

de Gelder, Beatrice ^{[3
]}

机构：

[1] ETH, BIWI, Zurich, Switzerland

[2] Katholieke Univ Leuven, VISICS, B-3001 Heverlee, Belgium

[3] Tilburg Univ, Cognit & Affect Neurosci Lab, Tilburg, Netherlands

来源：

NEURAL NETWORKS | 2008年 / 21卷 / 09期

关键词：

Body language; Categorical recognition; Emotion recognition; Biologically inspired model;

D O I：

10.1016/j.neunet.2008.05.003

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Research into the visual perception of human emotion has traditionally focused on the facial expression of emotions. Recently researchers have turned to the more challenging field of emotional body language, i.e. emotion expression through body pose and motion. In this work, we approach recognition of basic emotional categories from a Computational perspective. In keeping with recent Computational models of the visual cortex, we construct a biologically plausible hierarchy of neural detectors, which call discriminate seven basic emotional states from static views of associated body poses. The model is evaluated against human test subjects on a recent set of stimuli manufactured for research on emotional body language. (C) 2008 Elsevier Ltd. All rights reserved

引用

页码：1238 / 1246

页数：9

共 65 条

[1] Neural systems for recognizing emotion [J].