Fast image super-resolution algorithm based on multi-resolution dictionary learning and sparse representation
Fast image super-resolution algorithm based on multi-resolution dictionary learning and sparse representation
复制标题
DOI：
10.21629/jsee.2018.03.04
复制
发表时间：
2018-07
影响因子：
2.1
通讯作者：
Zhao Wei;Xiaofeng Bian;Huang Fang;Wang Jun;Abidi Mongi
中科院分区：
文献类型：
--
作者：
Zhao Wei;Xiaofeng Bian;Huang Fang;Wang Jun;Abidi Mongi
稀疏表示在过去十年中引起了广泛的关注，并在图像超分辨率（SR）方面表现出色。然而，目前许多图像超分辨率方法都面临着细节恢复和伪影抑制的矛盾。针对这一矛盾，提出了一种多分辨率字典学习（MRDL）模型，并给出了一种基于MRDL模型的快速单图像超分辨率方法。为了获得MRDL模型，我们首先使用我们提出的自适应补丁分区方法（APPM）提取多尺度补丁。APPM根据图像的细节丰富程度将图像划分为不同大小的块。然后，多分辨率字典对，其中包含各种分辨率的结构基元，可以从这些多尺度补丁训练。由于采用了MRDL策略，我们的SR算法不仅能够很好地恢复细节，具有更少的锯齿和噪声，而且还显著提高了计算效率。实验结果表明，该算法在评价指标和视觉感知方面优于其他SR方法。
: Sparse representation has attracted extensive attention and performed well on image super-resolution (SR) in the last decade. However, many current image SR methods face the contradiction of detail recovery and artifact suppression. We pro-pose a multi-resolution dictionary learning (MRDL) model to solve this contradiction, and give a fast single image SR method based on the MRDL model. To obtain the MRDL model, we fi rst extract multi-scale patches by using our proposed adaptive patch partition method (APPM). The APPM divides images into patches of different sizes according to their detail richness. Then, the multi-resolution dictionary pairs, which contain structural primitives of various resolutions, can be trained from these multi-scale patches. Owing to the MRDL strategy, our SR algorithm not only recovers details well, with less jag and noise, but also signi fi cantly improves the computational ef fi ciency. Experimental results validate that our algorithm performs better than other SR methods in evaluation metrics and visual perception.