Fast Regularization of Matrix-Valued Images
Abstract
Regularization of images with matrix-valued data is important in medical imaging, motion analysis and scene understanding. We propose a novel method for fast regularization of matrix group-valued images. Using the augmented Lagrangian framework we separate total- variation regularization of matrix-valued images into a regularization and a projection steps. Both steps are computationally efficient and easily parallelizable, allowing real-time regularization of matrix valued images on a graphic processing unit. We demonstrate the effectiveness of our method for smoothing several group-valued image types, with applications in directions diffusion, motion analysis from depth sensors, and DT-MRI denoising.
Cite
Text
Rosman et al. "Fast Regularization of Matrix-Valued Images." European Conference on Computer Vision, 2012. doi:10.1007/978-3-642-33712-3_13Markdown
[Rosman et al. "Fast Regularization of Matrix-Valued Images." European Conference on Computer Vision, 2012.](https://mlanthology.org/eccv/2012/rosman2012eccv-fast/) doi:10.1007/978-3-642-33712-3_13BibTeX
@inproceedings{rosman2012eccv-fast,
title = {{Fast Regularization of Matrix-Valued Images}},
author = {Rosman, Guy and Wang, Yu and Tai, Xue-Cheng and Kimmel, Ron and Bruckstein, Alfred M.},
booktitle = {European Conference on Computer Vision},
year = {2012},
pages = {173-186},
doi = {10.1007/978-3-642-33712-3_13},
url = {https://mlanthology.org/eccv/2012/rosman2012eccv-fast/}
}