Demosaicing based on the correlations between low-resolution images

  • Authors:
  • Hyun Mook Oh;Moon Gi Kang

  • Affiliations:
  • Yonsei University, Shinchon-Dong, Seodaemoon-Ku, Seoul, Korea;Yonsei University, Shinchon-Dong, Seodaemoon-Ku, Seoul, Korea

  • Venue:
  • SIP '07 Proceedings of the Ninth IASTED International Conference on Signal and Image Processing
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper, an edge directed demosaicing method based on the correlation between low-resolution images is proposed. A masaiced image from a Bayer pattern is regarded as combination of four up-sampled low-resolution images. In the beginning of the interpolation of each channel, the images are pre-interpolated toward the horizontal and vertical direction. The variances on the directionally interpolated image and on the difference domain is calculated based on the within- and cross-channel channel correlation. The missing values of each image is determined by using the variances. In the experimental results, proposed method shows improved result in the subjective and objective criterions.