hi, i'm trying to flip an image without any built in commands that is using for loops.

hi, i'm trying to flip an image without any built in commands that is using for loops. here is my code:

org_img=imread('Fig219(rose1024).tif');
subplot(4,3,1), imshow(org_img);
title('original image');
%---------------------------------------------------------------------------%
M=size(org_img,1);
N=size(org_img,2);

for i=1:M
    for j=1:N
        n_img(i,j)=org_img(i,N-j+1)
        subplot(4,3,2), imshow( n_img(i,j));
    end
end

on executing the above code the output goes out of bound can you please tell me what i'm doing wrong???

[Star Strider]

Define ‘flip’.

[Walter Roberson]

https://www.mathworks.com/matlabcentral/answers/38787-what-can-be-programmed-without-any-built-in-functions

[John BG]

diagonal flipping, kein problem:

A=imread('hccca.jpg');
A=A(:,:,1);
figure(1);imshow(A);
A=A(:,[end:-1:1],:);
figure(2);imshow(A([end:-1:1],:,:)');

.

.

John BG

<mailto:[e-mail] [e-mail]>

Answer 1

[John BG]

To saisps hmm

to flip images vertically or horizontally, there is no need for for loops and indeed any advanced function, just invert indices, look:

A=imread('netmap1.jpg')
figure(1);imshow(A);

figure(2);imshow(A([end:-1:1],:,:))

figure(3);imshow(A(:,[end:-1:1],:))

figures 1 2 3, from left to right

1: original

2: vertical flip (around X axis) and

3: horizontal flip (around Y axis)

.

.

if you find this answer useful would you please be so kind to mark my answer as Accepted Answer?

To any other reader, please if you find this answer of any help solving your question,

please click on the thumbs-up vote link,

thanks in advance

John BG

<mailto:[e-mail] [e-mail]>

[saisps hmm]

can an image be diagonally flipped by this technique of inverting indices?

[Walter Roberson]

Unfortunately, no, inverting indices is not enough for diagonal flipping.

[Walter Roberson]

I thought the question specifically required that for loops be used?

[saisps hmm]

no the question does not specifically required for loops. Main objective was to implement flipping without the use of built in commands. I was myself going for ''for loops''. so what method should be used for diagonally flipping an image?

[John BG]

Hi Saisps Hmm

Thanks for accepting my answer.

1.

Do you mean this?

A=imread('hccca.jpg');
A=A(:,:,1);
figure(1);imshow(A);
A=A(:,[end:-1:1],:);
figure(2);imshow(A([end:-1:1],:,:)');

.

.

Some people may consider that diagonal 'flipping' brings in 2 unknown parameters:

1. the angle.

2. what point: of the top left to low right diagonal, or for the matter, any parallel line to such main diagonal

To solve a similar problem I wrote the function imrotatex.m available from

https://uk.mathworks.com/matlabcentral/fileexchange/57492-rotate-image-around-arbitrary-point/content/imrotatex.m

please note that despite the rotated images look smaller, they aren't, the rotated image is 1:1. MATLAB zooms out and warns in the command window it has done so, in order to get a clear view of the transformed image.

I have attached imrotatext.m and example here

Regards

John BG

<mailto:[e-mail] [e-mail]>

[Walter Roberson]

No I have reread the question a few times and it does require for loops. "without any built in commands that is using for loops". Using for loops is the definition given to "without any built in commands". A solution that does not use for loops does not satisfy the question's definition of without built in commands.

If that is not the intended definition of without built in commands, then see the link I posted which goes into detail on what MATLAB considers to be built in commands.

[Jan]

The OP wrote: "no the question does not specifically required for loops".

A diagonal flipping can be achieved by permute easily:

img2 = permute(img, [2,1,3]);

[saisps hmm]

by diagonally flipping an image i mean the results to be as follows:

[saisps hmm]

sorry permute is a built in command don't want that :(

[Jan]

Ah, the old "no built-in" problem: after using Matlab for 20 years it is hard to stop thinking in terms of built-in functions. The image does not define the transformation uniquely, because the data are symmetric to the main diagonal. See my answer.

[Walter Roberson]

saisps hmm see the link I posted, https://www.mathworks.com/matlabcentral/answers/38787-what-can-be-programmed-without-any-built-in-functions .

Even indexing uses built-in commands.

Seriously: A(3,5) is not done by "just" indexing A at row 3, column 5: instead it is handled by calling

subsref(A, struct('type', '()', 'subs', {{3, 5}}) )

And this makes a difference in MATLAB, because it determines which subsref to call according to the data type of A. There are close to 50 different data types in MATLAB that define their own subsref, and you can define your own data types that use the () notation for special purposes.

There is almost nothing you can program in MATLAB without using a built-in function.

Answer 2

[Chad Greene]

If you want to flip it vertically (equivalent to flipud), you'd only need to loop through each row.

I = imread('coins.png');

Nrows = size(I,1); 

% Preallocate flipped version: 
Iflip = NaN(size(I)); 

for k = 1:Nrows
   Iflip(k,:) = I(Nrows-k+1,:); 
end

figure
subplot(1,2,1) 
image(I) 
axis image

subplot(1,2,2)
image(Iflip)
axis image
colormap gray

To flip it horizontally (equivalent to fliplr) you'd loop through columns in the same way.

[saisps hmm]

i understand the logic behind the for loop you used above but by using imshow() command the flipped window appears completely white

[Walter Roberson]

Instead of

Iflip = NaN(size(I));

use

Iflip = zeros(size(I), class(I));

[saisps hmm]

thanks, it worked!!!

[MUYIDEEN MOMOH]

your solution has solved my problem

Answer 3

[Jan]

Mirroring on the minor diagonal:

img  = rand(4, 4, 3);
imgT = cat(3, img(:,:,1).', img(:,:,2).', img(:,:,3).')
figure;
subplot(1,2,1);
image(img);
subplot(1,2,2);
image(imgT);

Or is the transpose and cat operation too built-in again? Note that the show code contains a built-in subsref and subsasgn also.

siz  = size(img);
imgT = zeros(siz([2,1,3]), class(img));
c    = siz(2) + 1;
for k = 1:siz(1)
  imgT(:, c - k, :) = img(k, end:-1:1, :);
end