基于边缘像素点分类和插值的图像放大新算法

A new image super-resolution algorithm based on classification of image edge pixels and interpolation

提出了一种超分辨率图像放大的新方法。根据低分辨率图像上的边缘信息,对初始的一个小图像块依据其对应的低分辨率图像上的边缘点进行分类,并根据此分类对图像块中某些像素点进行重新插值,得到放大的高分辨率图像块。由于对图像中的边缘点进行了特殊的处理,所提出的方法可以提高放大图像的边缘部分的清晰度,克服传统图像放大中图像过于平滑的缺点,对图像进行很好的放大。

In this paper, a new image super-resolution method for image processing is proposed. In this method, it first classifies the image edge pixels in the high resolution image patches into different categories according to the edge pixels in the low resolution image. Then, different interpolation methods are utilized for different pixels according to the decided categories. Since special processing for the image edge pixels is utilized, it can enhance the articulation of the high resolution image in the image edge areas, avoid the shortcomings of the traditional image super-resolution methods, and has a good super-resointion performance.