.png){kind=icon}
.png){kind=icon}
Return
Micro-scale searching algorithm for high-resolution image matting
.png){kind=icon}
PRE
AI
DOI：10.1007/s11042-023-17157-0.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Natural image matting based on pixel pair optimization is commonly employed during image post-processing. However, obtaining high-quality alpha mattes for high-resolution images via existing image matting methods is challenging as it typically requires considerable computational resources. In this paper, we design a novel optimization information transmission strategy that can be applied to images of different resolutions to improve the quality of the transmitted information required for evolutionary optimization. In addition, we propose a micro-scale searching matting algorithm, which allows us to obtain high-quality matting for high-resolution images with limited computational resources. To verify the applicability of the proposed algorithm for high-resolution images, experiments were conducted on the alpha matting benchmark dataset. Experimental results show that the proposed micro-scale searching matting algorithm can estimate high-quality alpha mattes without incurring excessive computational resources. Moreover, the proposed algorithm outperforms the state-of-the-art optimized matting algorithms when applied to high-resolution images.
Keywords:
Micro-scale searching algorithm
High-resolution image matting
Optimizing information transmission strategy
Computational resources
Journal
IF:
3
Papers:
1.9W
Citations:
3.2W