.png){kind=icon}
.png){kind=icon}
Return
Multiple Regressions based Image Super-resolution
.png){kind=icon}
PRE
AI
DOI：10.1007/s11042-019-7716-z.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The limitation of optical sensors set a challenge to acquire high resolution (HR) images. Previous sparse coding-based SR methods fail to reconstruct satisfied high resolution image due to three problems. First, sparse representation calculation is time consuming, which restricts its application in real-time systems. Second, sparse coding-based SR methods cannot represent diversity of patterns with one dictionary pair. Finally, it is supposed that the sparse representations of HR-LR patch pair are identical. However, the hypothesis cannot deal with all patterns. To address these problems, a multiple regressions based image super-resolution is proposed in this paper. First, to relax the hypothesis, the proposed method works on the assumption that the sparse representations of HR-LR patch pair are linear related. Secondly, training HR-LR patch pairs are departed into clusters. Then linear mappings is learned for each cluster. Finally, ridge regression is used to calculate the sparse representation. Experiments demonstrate that our method outperform some previous methods in objective and subjective evaluation. Additionally, our method is less computational complexity.
Keywords:
Super-resolution
Sparse coding
Ridge Regression
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
3
Papers:
1.9W
Citations:
3.2W