一种新的基于感知字典的稀疏图像重建算法研究  

A New Algorithm for Sparse Image Reconstruction Based on Sensing Dictionary

作　　者：陈瑞瑞 李爽 

机构地区：[1]郑州工业应用技术学院信息工程学院,河南郑州451100

出　　处：《洛阳师范学院学报》2017年第8期32-35,共4页Journal of Luoyang Normal University

基　　金：河南省科技攻关计划项目(0721002210032)

摘　　要：针对已有图像重建算法分辨率低、需要大量计算的问题,本文提出了一种基于感知字典和数据自适应性的稀疏重建算法.首先,针对图像的数据结构,对样本数据进行超完备字典的训练,继而通过针对性的字典对图像进行稀疏重建.同时,为进一步改善算法的重建性能,并充分利用图像的有效信息,本文构造了数据自适应的感知字典.实验表明,该算法在不影响图像重建精确度的前提下可以减少计算复杂度,并具有良好的鲁棒性和较高的效率.For existing image reconstruction algorithm, the resolution is low and the computation load is high. A sparse reconstruction algorithm based on sensing dictionary and data adaptability is proposed. First, for the data structure of the image, the sample data is trained with super-complete dictionary, and then sparse reconstruction of the image is implemented through targeted dictionary. At the same time, in order to further improve the reeonstruc- tion performance of the said algorithm and make full use of the effective information of the image, a data adaptive sensing dictionary is established. Experiments indicate that the algorithm can reduce the computational complexity on the premise that image reconstruction accuracy is not affected, and has good efficiency and robustness.

关 键 词：感知字典 稀疏重建 自适应 数据结构 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]