计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (2): 169-172.DOI: 10.3778/j.issn.1002-8331.2011.02.051

• 图形、图像、模式识别 • 上一篇    下一篇

蚁群算法在呼吸信号情感识别中的应用研究

林时来,刘光远,张慧玲   

  1. 西南大学 计算机与信息科学学院,重庆 400715
  • 收稿日期:2010-05-04 修回日期:2010-07-12 出版日期:2011-01-11 发布日期:2011-01-11
  • 通讯作者: 林时来

Application of ACO algorithm to emotion recognition research based on RSP signal

LIN Shilai,LIU Guangyuan,ZHANG Huiling   

  1. College of Computer and Information Science,Southwest University,Chongqing 400715,China
  • Received:2010-05-04 Revised:2010-07-12 Online:2011-01-11 Published:2011-01-11
  • Contact: LIN Shilai

摘要: 针对生理信号的情感识别问题,将蚁群优化算法用于呼吸信号(RSP)特征选择,并采用自适应的适应度参数值、变异策略和临近位置交换策略对算法进行改进,使用Fisher进行情感分类,获得了较高的识别率和有效特征组合。采集了212个被试6种不同情感(高兴、惊奇、厌恶、悲伤、愤怒、恐惧)的呼吸信号数据进行仿真实验,识别效果最好的是高兴情感,最好识别率达到了92.06%,平均识别率达到了84.43%。实验仿真结果表明,将蚁群优化算法引入基于呼吸信号的情感识别研究是可行的。

关键词: 情感识别, 蚁群算法, 呼吸信号, 特征选择

Abstract: This study is on the emotion recognition method based on respiration(RSP) signals.Firstly,six emotional(Happiness,Surprise,Disgust,Grief,Anger,Fear) RSP signals of 212 subjects are de-noised and original features are extracted with wavelet transform method.Then,feature selection is done using the ant colony optimization(referred to as ACO) algorithm.Results show that after ACO algorithm is optimized,emotion recognition of physiological signal has achieved good results.Happiness emotion has a best recognition rate of 92.06%,the corresponding average recognition rate achieves 84.43%.

Key words: emotion recognition, Ant Colony Optimization(ACO), RSP signal, feature selection

中图分类号: