ML Anthology

Radiometric Calibration from a Single Image

Stephen Lin, Jinwei Gu, Shuntaro Yamazaki, Heung-Yeung Shum
CVPR 2004 pp. 938-945
doi:10.1109/CVPR.2004.197 /cvpr/2004/lin2004cvpr-radiometric/

Abstract

Photometric methods in computer vision require calibration of the camera's radiometric response, and previous works have addressed this problem using multiple registered images captured under different camera exposure settings. In many instances, such an image set is not available, so we propose a method that performs radiometric calibration from only a single image, based on measured RGB distributions at color edges. This technique automatically selects appropriate edge information for processing, and employs a Bayesian approach to compute the calibration. Extensive experimentation has shown that accurate calibration results can be obtained using only a single input image.

CVPR Semantic Scholar

Cite

Text

Lin et al. "Radiometric Calibration from a Single Image." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2004. doi:10.1109/CVPR.2004.197

Markdown

[Lin et al. "Radiometric Calibration from a Single Image." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2004.](https://mlanthology.org/cvpr/2004/lin2004cvpr-radiometric/) doi:10.1109/CVPR.2004.197

BibTeX

@inproceedings{lin2004cvpr-radiometric,
  title     = {{Radiometric Calibration from a Single Image}},
  author    = {Lin, Stephen and Gu, Jinwei and Yamazaki, Shuntaro and Shum, Heung-Yeung},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year      = {2004},
  pages     = {938-945},
  doi       = {10.1109/CVPR.2004.197},
  url       = {https://mlanthology.org/cvpr/2004/lin2004cvpr-radiometric/}
}