Image super-resolution based on shock filter and non-local means
Image super-resolution based on shock filter and non-local means
复制标题
DOI：
10.1109/apsipa.2014.7041613
复制
发表时间：
2014-02
期刊：
影响因子：
--
通讯作者：
Taichi Yoshida;M. Ikehara
中科院分区：
文献类型：
--
作者：
Taichi Yoshida;M. Ikehara
本文提出了一种基于冲击滤波和非局部均值的超分辨率方法。超分辨率放大后的图像边缘清晰、文字清晰、纹理细腻。该方法通过冲击滤波逐步提高轮廓的清晰度，并通过非局部方法有效地保持轮廓的结构和纹理。在一些模拟中，所提出的方法表现出更好的结果比传统的和可比的结果，国家的最先进的超分辨率方法，需要高的计算复杂度。所提出的方法需要低的计算复杂度，因为它可以通过简单的迭代过程来实现。
In this paper, we propose a super-resolution method based on the shock filter and non-local means. The super-resolution generates the enlarged images which have sharp edges, clear characters and fine textures. The proposed method gradually improves the sharpness of contours by the shock filter and effectively maintains the structures and textures by the nonlocal means. In some simulations, the proposed method shows better results than conventional ones and comparable results to the state-of-the-art super-resolution method which requires high computational complexities. The proposed method requires low computational complexities, because it can be realized by simple iterative processes.