본문 바로가기
  • Home

Similarity Enhancement of Face Recognition Performance between Real Faces and Facial Videos through Adjustment of a Display Device

  • Journal of Knowledge Information Technology and Systems
  • Abbr : JKITS
  • 2016, 11(2), pp.197-204
  • Publisher : Korea Knowledge Information Technology Society
  • Research Area : Interdisciplinary Studies > Interdisciplinary Research
  • Published : April 30, 2016

Mi Young Cho 1 Young-sook Jeong 1 Byung-Tae Chun 2

1한국전자통신연구원
2한경대학교

Accredited

ABSTRACT

As face recognition technology, unlike other biometric methods, is more direct, user friendly, and convenient, it is applicable to many areas including security systems and service robot. However, the performance of face recognition systems in a real service environment does not always satisfy the expectations of users, because the performance of face recognition systems is influenced by the user's pose, expression change and lighting in the real environment. So, there is the performance gap between a real service and laboratory environment. To reduce the gap, there are many approaches using such as an algorithm, a mannequin and a high-definition photograph. Unfortunately, it has a limitation to construct a test environment in which the various situations faced by a user and an face recognition system are created similar to those encountered in real environments. Recently, with the development of the display device and capture devices, the performance evaluation method using the high-quality face video output from the display device in place of the actual human face to are proposed. In this paper, we demonstrate the similarity enhancement of face recognition performance between real face and facial videos from display device through gamma adjustment.

Citation status

* References for papers published after 2023 are currently being built.