Application of Bayesian Belief Network in Reliable Analysis for Video Deinterlacing

  • Authors:
  • Gwanggil Jeon;R. Falcon;Donghyung Kim;Rokkyu Lee;Jechang Jeong

  • Affiliations:
  • Hanyang Univ., Seoul;-;-;-;-

  • Venue:
  • IEEE Transactions on Consumer Electronics
  • Year:
  • 2008

Quantified Score

Hi-index 0.43

Visualization

Abstract

In this paper, we illustrate that Bayesian networks (BNs), which are also known as belief networks, are well-suited for image processing. We provide case studies on video deinterlacing methods. The proposed efforts at modeling weight measuring process involved in weight assignment of conventional deinterlacing methods that are commonly used for industrial world. Using probabilistic BNs, the system determines the weights and interpolates the missing pixels robustly. The results of empirical trial show that the proposed system can deal successfully with several types of images containing motion or detail.