A Statistical Approach for Recognizing Emotion from Dance Sequence

  • Park, Han-Hoon (Department of Electrical and Computer Engineering Hanyang University) ;
  • Park, Jong-Il (Department of Electrical and Computer Engineering Hanyang University) ;
  • Kim, Un-Mi (Department of Dance, Hanyang University) ;
  • Woontack Woo (U-VR Lab., K-JIST)
  • 발행 : 2002.07.01

초록

We propose a simple method that can recognize human emotion from monocular dance image sequences. The method only exploits the information within image sequences and does not require cumbersome attachments like sensors. This makes the method a simple, human-friendly one. Moreover, the method is more robust and efficient by taking into account the statistical property of image sequences based on PCA (Principal Component Analysis). The correct recognition rate in real-time is about 75% in a variety of experiments.

키워드