提出了一种线性和非线性相结合的图像缩小方法。首先由近邻取样方法和邻域平均相结合的方法产生一个中间图像1,利用二元Newton—Thiele型向量连分式建立有理插值曲面;然后对插值曲面进行重新采样产生一个中间图像2;最后对这两幅中间图像进行加权求和,得到最终的图像。实验表明,该方法比常规方法产生的缩小图像效果好。
A new method of image zooming was presented, which combined non-linear method with linear one. First a transition image 1st was generated by means of neighborhood pixels, and then, a transition image 2 was generated by resampling the surface of Newton-Thiele' s vector valued interpolation function. At last, the weighted average of the two transition images resuited in the final image author wished to get. The experimental results show that the method is superior to those of classical interpolation schemes.