Minimum variance method to obtain the best shot in video and its effectiveness for face recognition

Kazuo Ohzeki, Ryota Aoyama, Yutaka Hirakawa

Research output: Contribution to journalArticle


This paper describes a face recognition algorithm using feature points of face parts, which is classified as a feature-based method. As recognition performance depends on the combination of extracted feature points, we utilize all reliable feature points effectively. From moving video input, well-conditioned face images with a frontal direction and without facial expression are extracted. To select such well-conditioned images, an iteratively minimizing variance method is used with variable input face images. This iteration drastically brings convergence to the minimum variance of 1 for a quarter to an eighth of all data, which proves to take the frontal image in 0.27 second from video at most. The proposed system using six statistic values realizes 98.3% as an authentication rate.

Original languageEnglish
Pages (from-to)29-44
Number of pages16
JournalInternational Journal of Computer Science and Applications
Issue number2
Publication statusPublished - 2016



  • Distance
  • Distribution
  • Face recognition
  • Feature point

ASJC Scopus subject areas

  • Computer Science Applications

Cite this