Image interpolation in MATLAB involves interp1

Question

currently I'm struggling to create a solution to resize an image(colour and greyscale) using interp1() in MATLAB. I've looked everywhere but the only solution I found is to use interp2() which is something I can't. I have been trying and researching for almost 3 days now, but nothing help. thank you in advance.

I understand that it is easier for me to use interp2 but in this question, i can only use interp1. — thanh thai, Jun 01 '17 at 08:34
I have been looking at this [post](https://stackoverflow.com/questions/6183155/resizing-an-image-in-matlab) and trying to refer to it but something is still wrong or missing. — thanh thai, Jun 01 '17 at 08:37
If you have a RGB image read as I, you can try using: I = I(1:2:end,1:2:end,3) etc..... — Siva Srinivas Kolukula, Jun 01 '17 at 08:42
@SivaSrinivasKolukula yes I have to work with both RGB and greyscale but I still don't understand what you mean. Can u explain that further? thanks. — thanh thai, Jun 01 '17 at 08:48

score 0 · Answer 1 · answered Jun 01 '17 at 08:53

0

YOu may skip few rows and columns, if you are against interpolation.

I = imread('peppers.png') ;   
size(I)
I1 = I(1:2:end,1:2:end,:) ;
size(I1)

And also, you can have a look into imresize.

I2 = imresize(I,[200 200]) ;

answered Jun 01 '17 at 08:53

Siva Srinivas Kolukula

1,251
1
7
14

I must use interp1() in this problem and imresize() must not be used in this case. thanks in advance. – thanh thai Jun 01 '17 at 08:59

score 0 · Accepted Answer · answered Jun 01 '17 at 10:39

As you are strict on using interp1..you may proceed as follows. Do interpolation based on global indices.

I = imread('peppers.png') ;
[nx,ny,t] = size(I) ;
%%
dx = 4 ; dy = 4 ;   % see to it that dx and dy are multiples of nx and ny
%% Global indices 
idx = 1:nx*ny ;
% for interpolation 
idxi = 1:dx:nx ; idyi = 1:dy:ny ;
[I1,J1] = meshgrid(idxi,idyi) ;
idxi = sub2ind([nx,ny],I1,J1)' ;
%%
nxi = nx/dx ; nyi = ny/dy ;
I1 = zeros(nxi,nyi,t) ;
%%
for i = 1:t
    C = I(:,:,i) ;
    temp = interp1(idx,double(C(:)),idxi(:)) ;
    I1(:,:,i) = reshape(temp,nxi,nyi) ;
end
I1 = uint8(I1) ;
imshow(I1)
size(I)
size(I1)

thank you so much @siva props to you man :) i will look into this. — thanh thai, Jun 01 '17 at 11:21

Image interpolation in MATLAB involves interp1

2 Answers2