Abstract: A method of image interpolation enlargement based on Cycle-Spinning is proposed in wavelet domain.Cycle-Spinning has been shown to be effective in image denoising in order to avoid Gibbs-like artifacts.Using the given lower resolution image，the original high resolution image is approximated using zero padding of high-frequency sub-bands followed by inverse wavelet transform.Cycle-Spinning technique is then adopted to improve the resolution of the original high resolution image，and the edge.Simulation results show that proposed method is fast，simple and effective compared with several traditional or emerged improved algorithms，and the higher quality of an enlarged image can be obtained both on edges and in smooth area.

Key words: Cycle-Spinning, interpolation, enlargement, wavelet transform

摘要： 提出一种基于Cycle-Spinning的小波域图像插值放大算法。Cycle-Spinning技术已证明是小波去噪中消除伪Gibbs效应的有效方法，将低分辨率图像作为高分辨率图像的低频子带，将高频系数置零，再进行反变换，得到高分辨率图像的初始估计。在此基础上，利用Cycle-Spinning技术处理得到高分辨率图像，很好地消除了边缘锯齿现象。经实验仿真，并与传统及改进插值算法比较，提出的算法简单、快速、有效，使放大后的图像无论在边缘处还是平滑区域均能达到理想的效果。

关键词: Cycle-Spinning, 内插, 放大, 小波变换