基于Contourlet的图像超分辨率重构算法

Image Super Resolution Reconstruction Algorithm Based on Contourlet

提出一种新颖的基于Contourlet的图像超分辨率重构算法,该算法对图像进行Contourlet分解后得到图像的各个方向的细节信息分块处理,根据每个小块在已有的图像库中寻找最佳匹配,恢复出低分辨率损失的系数,然后进行Contourlet重构,得到最终结果。通过采用图像各个方向的信息,重构出的图像更忠实于原图。实验分析表明,该算法能有效地实现图像超分辨重构,得到的结果更加优化。

Proposes a new method of super resolution based on Contourlet.Uses Contourlet to decompose the image to get information in different directions of image,cuts the image to little patch and finds the best match in the dictionary,thus can get the coefficients that we have lost.With these coefficients we can use Contourlet to compose the original image that we are excepting.The final image is more faithful to the original one because we use all directions other than four directions,it＇s much better than other methods.