How were these two plots created?

I've got two plots from scientific papers that I can't seem to recreate. I would apprechiate if someone could tell me which plot command was used.

The first one is a color-coded representation of values in a matrix (or at least I assume it is), where the number that is represented is also visible in the image.

I've tried playing around with the imagesc() and the pcolor() commands, but none of them allowed me to display the numeric value represented by the color.

My second enigmatic plot is the result of a 2D Fourier transformation:

I'm pretty sure it was created via imagesc(), but I'd like to know how the axis midpoint was moved into the center of the image and how the axis cross was created.

When I try

iamgesc(imfftnorm);
colormap(jet)
ax = gca;
ax.XAxisLocation = 'origin';
ax.YAxisLocation = 'origin';

I get this instead:

Different shape aside (I think that's because I was only using a screenshot of his data for the Fourier transformation), I can't get the middle of the axes moved to the center of the image, and I don't know how that cross can be displayed.

If anyone has any idea about these two, I'd be very grateful.

Answer 1

[Star Strider]

The top plot image appears to be a heatmap.

The lower plot image is most likely a surf plot using the view function (specifically view(0,90)) to see it from the top.

That’s my best guess for the lower one, anyway.

[Korbi]

Thanks, Heatmap is exactly what I was looking for.

The Fourier transformation still doesn't look like in that image, but I suppose I can just manually put in the cross.

[Star Strider]

As always, my pleasure!

The ‘cross’ pattern of the fft2 plot is likely the result of the resolution of the original signal. It will likely be more apparent with a different spatial resolution.

[Korbi]

Oh, I didn't mean the "cross" pattern, I meant the axes, the lines that go through the center point of (fx=0,fy=0). The patterns around the high points are an artifact of Fourier transformation without windowing.

[Star Strider]

You can use vectors (of the appropriate lengths) or matrices (of the appropriate sizes) for the ‘x’ and ‘y’ surf plot coordinates. Use the linspace function to create the vectors going from -0.5 to +0.5 with the specified lengths. It may be necessary to specifically define the tick values (and perhaps labels) in order to get the plot to appear the say you want it to. There are several links under the See Also section of the axes documentation that will make that task easier.

I’m not certain exactly what you want to do, however with a few details I can probably help with that.

[Korbi]

Thanks, I think that'll get me the axes. As for the coordinate cross, I'll probably put it in via powerpoint, like that, except with more care:

It's pretty superfluous, I only wanted to put it in because my professor pointed at a plot like that and said he wanted it presented like in that picture.

[Star Strider]

As always, my pleasure!

I had to take a closer look to see the black axes in the centre. I didn’t realise they were there until you mentioned it.

You can add the cross programmatically with:

hold on
plot3([1 1]*mean(xlim), ylim, [1 1]*max(zlim), '-k')
plot3(xlim, [1 1]*mean(ylim), [1 1]*max(zlim), '-k')
hold off

the full code being:

figure
surf( FFT2_matrix , 'EdgeColor','none')
hold on
plot3([1 1]*mean(xlim), ylim, [1 1]*max(zlim), '-k')
plot3(xlim, [1 1]*mean(ylim), [1 1]*max(zlim), '-k')
hold off
colormap(jet)
view(0,90)

That worked when I tried it on an example surf plot.

.

[Korbi]

Hah, how did I not think of that? Sometimes the easiest answer is the best answer.

Thanks, I thought that maybe the original had used some setting to move the x- and y-axis to the middle of the screen.

[Star Strider]

As always, my pleasure!

I thought these:

ax = gca;
ax.XAxisLocation = 'origin';
ax.YAxisLocation = 'origin';

would work, however it may be necessary to add:

ax.ZAxisLocation = 'origin';

as well. (I didn’t consider that possibility, and I didn’t test it since creating them manually worked.)