Colorization by Patch-Based Local Low-Rank Matrix Completion

Abstract

Colorization aims at recovering the original color of a monochrome image from only a few color pixels. A state-of-the-art approach is based on matrix completion, which assumes that the target color image is low-rank. However, this low-rank assumption is often invalid on natural images. In this paper, we propose a patch-based approach that divides the image into patches and then imposes a low-rank structure only on groups of similar patches. Each local matrix completion problem is solved by an accelerated version of alternating direction method of multipliers (ADMM), and each AD-MM subproblem is solved efficiently by divide-and-conquer. Experiments on a number of benchmark images demonstrate that the proposed method outperforms existing approaches.

Cite

Text

Yao and Kwok. "Colorization by Patch-Based Local Low-Rank Matrix Completion." AAAI Conference on Artificial Intelligence, 2015. doi:10.1609/AAAI.V29I1.9479

Markdown

[Yao and Kwok. "Colorization by Patch-Based Local Low-Rank Matrix Completion." AAAI Conference on Artificial Intelligence, 2015.](https://mlanthology.org/aaai/2015/yao2015aaai-colorization/) doi:10.1609/AAAI.V29I1.9479

BibTeX

@inproceedings{yao2015aaai-colorization,
  title     = {{Colorization by Patch-Based Local Low-Rank Matrix Completion}},
  author    = {Yao, Quanming and Kwok, James T.},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year      = {2015},
  pages     = {1959-1965},
  doi       = {10.1609/AAAI.V29I1.9479},
  url       = {https://mlanthology.org/aaai/2015/yao2015aaai-colorization/}
}