Zooming by Linear Interpolation
Linear interpolation is a FOH. For eg. For a 2 2 matrix
for ; ![](8_12_clip_image008.gif)
for , ![](8_12_clip_image014.gif)
Linear interpolation along columns gives
correct expression ![](8_12_clip_image018.gif)
![](8_12_clip_image020.gif)
for , ![](8_12_clip_image022.gif)
(corrected version )
for , )
(assume i/p image is zero outside )
![](matx.gif)
Problem: Show the above result can also be obtained by convolving the 2M× 2N zero interlaced images with the array
![](8_12_clip_image002_0000.gif)
whose origin ( m=0, n=0 ) is at the centre of the array (boxed element). Higher order interpolation is possible by padding each row and column of input image by p rows and p column of zeros and convolving it p times with H.
|