Set xaxis of imagesc with scaled values

Hi everybody,

I have a matrix(20x400) and I am plotting it with imagesc where y axis having 20 values and xaxis having 400 values.

However, I would like to know how can I scale this xaxis 400 to intervals like between 0:20 = 1, 20:40 = 2 until 380:400 = 20; and setting x axis of imagesc in 0-20 scale with the values of 0-20.

I hope it is clear what I am intended to do.

[DGM]

Are you just wanting to set the spacing for the ticks, or are you actually wanting a nonuniform spatial scaling?

Answer 1

[Adam Danz]

Specify x and y values using imagesc(x,y,C)

I think this is what you're looking for

C = rand(20,400); % Replace this with your matrix
x = linspace(1,height(C),width(C)); 
y = 1:height(C); 
imagesc(x,y,C)
axis equal % optional
axis tight % optional

Alternatively, use imresize

figure
I = imresize(C,[20,20]); 
imagesc(I)
axis equal
axis tight

[Ahmet Hakan UYANIK]

Thank you for the answer, it works like that way.

Also if I want to plot Nan values in the data as white color, I have been using pcolor as folllows:

C = replace_matrix; % Replace this with your matrix

x = 1:20;

y = 1:height(C);

imagesc(x,y,C)

[nr,nc] = size(C);

pcolor([C nan(nr,1); nan(1,nc+1)]);

shading flat;

However, it changes the axis again, how can I put white color to NaN values without changing the axis scale?

[Ahmet Hakan UYANIK]

okay I did

imagesc(x,y,C,'AlphaData',~isnan(C))

and it works like this.

[Ahmet Hakan UYANIK]

One more doubt, when I scale the axis as stated, image is horizantally disturbed(below figure). However, it was not the case before(figure above). Do you know how to handle with it?

[Adam Danz]

My solution scales the x-values 1:400 to 1:20 so that there are still 400 values. This vertically stretches each image pixel because now there are 20 x values for each y value.

If you want the pixels to remain square, you can resize the axes so that the ratio of the width and height correspond the ratio of the data as shown below. But as you see, this results in a short and long axes which is probably not what you're looking for.

C = rand(20,400); % Replace this with your matrix
x = linspace(1,height(C),width(C)); 
y = 1:height(C); 
imagesc(x,y,C)

ax = gca(); 
ax.Position(4) = ax.Position(3) * height(C)/width(C);

Another approach is to use imresize to rescale your image.

figure
I = imresize(C,[20,20]); 
imagesc(I)

[Ahmet Hakan UYANIK]

Excellent, thanks!

[Adam Danz]

Glad it worked out. I'll add the imresize suggestion to the main answer for better visibility.