摘要
本文提出了一种针对彩色图像的区域指导的插值算法.该方法首先从低分辨率图像中计算象素沿梯度方向的1、2阶方向导数,然后把彩色图像划分为常数、线性和非线性区域,并对不同的区域施用不同复杂度的滤波器.实验结果表明,与传统的双线性插值相比,该方法可以明显改进插值图像的质量;重建所需的时间也比单纯使用复杂滤波器的时间减少很多,并且可以实现彩色图像的任意放大.
The paper proposes an area-directed interpolation algorithm for color images. The basic idea is to first estimate first and second direction derivatives around interpolated pixels along the gradient directions from a low-resolution image and then identify constant zones, linear zones, and non-linear zones based on these direction derivatives in color image. Thus different filters are applied only to zones that require it. Experiment results demonstrate that our new interpolation algorithm substantially improves the quality of the interpolated images over conventional bilinear interpolation and possesses reconstruction cost less than pure complex filters. In addition, the proposed method can be used irre- spective of the magnification factor （MF）.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2006年第1期35-39,共5页
Acta Electronica Sinica
基金
安徽省自然科学基金(01042202)
关键词
彩色空间
边界特征
自适应插值
图像重建
color space
edge feature
adaptive interpolation
image reconstruction
