摘要
针对高斯模板用于图像旋转时,使图像产生模糊的问题,提出细粒度动态单像素模板算法。根据图像小邻域局部相关性原理,将当前像素分割为5×5的子像素,对每一个子像素求得一个模板。用当前像素的这25个模板之一代替高斯模板进行图像旋转,消除了高斯模板旋转图像时产生模糊的问题。该算法旋转图像的质量与双线性插值算法相近,且具有更高的计算性能。
A fine granularity dynamic single pixel template algorithm is proposed to resolve blur problem of image rotated by Gaussian template. According to the principle of local small neighborhood relativity, the present pixel is split into sub-pixel of 5×5. Templates for each sub-pixel are designed. The blur problem of rotated image is resolved when image is rotated using one of the twenty-five templates while not the Gaussian template. The quality of rotated image using the algorithm is similar to that of the bilinear interpolation algorithm, but the performance is higher than that of the bilinear interpolation.
出处
《计算机工程与设计》
CSCD
北大核心
2008年第16期4267-4270,共4页
Computer Engineering and Design
基金
广东省科技厅科研基金项目(2003C32401)
关键词
单像素模板
像素分割
高斯模板
双线性插值
图像旋转
single pixel template
pixel segmentation
Gaussian template
bilinear interpolation
image rotation
