摘要
为获得清晰光滑的图像轮廓,提出一种基于泰勒展开式的图像插值方法。根据靠近边缘中心一侧的已知像素,利用泰勒展开式估计待插值像素,减小插值图像边缘的宽度,并增加边缘斜坡坡度,从而获得清晰的插值图像边缘。实验结果表明,该方法简单且易于实现,与逆梯度加权方法相比,能获得更清晰的图像边缘。
To obtain crisp and smooth image contour of interpolated image, an image interpolation method based on Tyler expansion is proposed. The interpolated pixel in a lateral near the center of edge is approximated by the pixel in this lateral by Tyler expansion. It reduces width of edge and increases slope, and produces scrip edges. Experimental results show that the advantage of this method is simple and easy to implementation. The method can obtain crisper border compared with inverse gradient weighting method.
出处
《计算机工程》
CAS
CSCD
2012年第13期202-204,共3页
Computer Engineering
基金
重庆市自然科学基金资助项目(cstcjjA40012)
关键词
图像插值
图像缩放
泰勒展开式
逆梯度权
双线性插值
image interpolation
image zooming
Taylor expansion
inverse gradient weight
bilinear interpolation
作者简介
詹毅（1971-），男，副教授、博士，主研方向：图像处理，模式识别E-mail：zhanyimeng@yahoo．com．cn
