基于自适应字典稀疏表示的超分辨率算法

基于学习的超分辨率算法通过一组训练样例来学习一个字典,并从该字典中合成低分辨率图像中丢失的高频信息,最终得到相应的高分辨率图像。介绍了几种常用的基于学习的超分辨率算法,并提出了一种新的算法:基于自适应字典稀疏表示的超分辨率算法。实验结果表明,该方法在主观与客观上均具有较好的重建效果。

Sparse representation over adaptive dictionaries for super-resolution

Learning-based super-resolution algorithm tries to learn a dictionary from a set of training images,and synthesizes the lost high-frequency information of an input low-resolution image to acquire the corresponding high-resolution image.This paper introduces the related work of learning-based super-resolution algorithms and proposes a new method: sparse representation over adaptive dictionaries for super-resolution.The experimental results show that the proposed algorithm has a better reconstruction quality in both objective and subjective comparisons.