基于二元树复小波的图像超分辨率内插  

Image Super-resolution Interpolation Based on Dual-tree Complex Wavelet Transformation

作　　者：袁小华[1] 刘纯平[2] 高秀梅[3] 夏德深[1] 

机构地区：[1]南京理工大学计算系603教研室 [2]苏州大学计算机科学与技术学院,苏州215006 [3]淮阴师范学院计算机科学系,淮安223001

出　　处：《计算机工程》2006年第8期26-28,共3页Computer Engineering

基　　金：香港特区政府研究资助局资助项目(CUHK/4180/01E);江苏省教育厅自然科学基金资助项目(04KJD520037)

摘　　要：在Nyugen N等的图像超分辨率小波内插的算法框架下,用二元树复小波变换代替实离散小波变换,使图像超分辨率小波内插能得到更好的高分辨率图像的拟合部分和更多方向的细节。算法分析和实验结果表明了该方法是对N.Nyugen等的图像超分辨率小波内插框架的进一步改进。Under the framework of image super-resolution wavelet interpolation proposed by Nyugen N and et al, by taking advantage of characters such as shift-invariance, better directive selection and effective computation of dual tree complex wavelet, a Q-shift version of dual tree complex wavelet is introduced, so as to obtain a better interpolated approximate image with much less artifacts and more directions of details. Algorithm analysis and results of tests on synthetic image series are also presented to show the improvements to the works of Nyugen N and et al.

关 键 词：图像超分辨率小波内插 交错采样 二元树复小波变换 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]