Aggregating color and absolute difference for CFA interpolation

Yifan LiuICCV 2022 · DOI

Abstract

Reconstructing images from raw pictures taken by digital cameras has always been a hot topic in the computer vision community, which aims to transform the intensity of light captured by the single charge-coupled device(CCD) to pixel color information of all channels. As a commonly used technique, interpolation has been proven to be an effective solution for this challenging task, but its performance is still limited especially when dealing with the edges of images. In this paper, we propose a three-stage color reconstruction solution for digital cameras, which allocates weights on neighboring pixels using aggregated color difference on stage one and aggregated absolute difference on the last two stages. Experimental results show that our proposed interpolation method outperforms existing interpolation methods on most images in both subjective and objective measurements.
BibTeX
@inproceedings{liu2022aggregating,
  author = {Yifan Liu},
  title = {{Aggregating color and absolute difference for CFA interpolation}},
  booktitle = {IEEE/CVF International Conference on Computer Vision},
  year = {2022},
  doi = {10.1109/cvidliccea56201.2022.9824744},
}

Loading related content...

Figure