Journal
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
Volume 27, Issue 9, Pages 1856-1868Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TCSVT.2016.2555779
Keywords
Color-to-gray conversion; RGB2GRAY; semiparametric; subspace modeling; two-order polynomial model
Categories
Funding
- National Natural Science Foundation of China [61261010, 61362001, 61503176, 51165033]
- International Scientific and Technological Cooperation Projects of Jiangxi Province [20141BDH80001]
- Young Scientists Training Plan of Jiangxi Province [20142BCB23001]
- Jiangxi Advanced Projects for Post-Doctoral Research Funds [2014KY02]
- International Post-Doctoral Exchange Fellowship Program
Ask authors/readers for more resources
While the RGB2GRAY conversion with fixed parameters is a classical and widely used tool for image decolorization, recent studies showed that adapting weighting parameters in a two-order multivariance polynomial model has great potential to improve the conversion ability. In this paper, by viewing the two-order model as the sum of three subspaces, it is observed that the first subspace in the two-order model has the dominating importance and the second and the third subspace can be seen as refinement. Therefore, we present a semiparametric strategy to take advantage of both the RGB2GRAY and the two-order models. In the proposed method, the RGB2GRAY result on the first subspace is treated as an immediate grayed image, and then the parameters in the second and the third subspace are optimized. Experimental results show that the proposed approach is comparable to other state-of-the-art algorithms in both quantitative evaluation and visual quality, especially for images with abundant colors and patterns. This algorithm also exhibits good resistance to noise. In addition, instead of the color contrast preserving ratio using the first-order gradient for decolorization quality metric, the color contrast correlation preserving ratio utilizing the second-order gradient is calculated as a new perceptual quality metric.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available