作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2012, Vol. 38 ›› Issue (20): 191-194. doi: 10.3969/j.issn.1000-3428.2012.20.049

• 图形图像处理 • 上一篇    下一篇

基于L1/2正则化的超分辨率图像重建算法

王 欢,王永革   

  1. (北京航空航天大学数学与系统科学学院,北京 100191)
  • 收稿日期:2011-12-31 修回日期:2012-02-26 出版日期:2012-10-20 发布日期:2012-10-17
  • 作者简介:王 欢(1988-),女,硕士研究生,主研方向:图像处理;王永革,副教授
  • 基金资助:
    国家自然科学基金资助项目(10801007);国家“973”计划基金资助项目(2010CB731900)

Super-resolution Image Reconstruction Algorithm Based on L1/2 Regularization

WANG Huan, WANG Yong-ge   

  1. (School of Mathematics and Systems Science, Beihang University, Beijing 100191, China)
  • Received:2011-12-31 Revised:2012-02-26 Online:2012-10-20 Published:2012-10-17

摘要: 为提高图像重建质量,研究超分辨率图像重建技术与稀疏表示理论,提出一种基于L1/2正则化的超分辨率图像重建算法。将L1/2正则化理论运用到字典学习中,利用学习得到的字典重建高分辨率图像。实验结果表明,该算法的图像重建效果优于基于L1正则化的超分辨率图像重建算法。

关键词: L1/2正则化, 稀疏表示, 超分辨率图像重建, K-SVD算法, 字典学习, 训练样本

Abstract: In order to improve the image reconstruction quality, by studying the super-resolution image reconstruction technology and the theory of sparse representation, this paper proposes a super-resolution image reconstruction algorithm based on L1/2 regularization. It applies L1/2 regularization into dictionary learning, and reconstructs super-resolution images using learned dictionaries. Experimental results show that the reconstruction results in this paper are better than the results of super-resolution image reconstruction algorithm based on L1 regularization.

Key words: L1/2 regularization, sparse representation, super-resolution image reconstruction, K-SVD algorithm, dictionary learning, training sample

中图分类号: