# Impulse response of the system image processing

In case of the signal, when we plot an impulse response we have our x-axis - samples (or time), y-axis - amplitude, but in case I work with image processing?

x-axis is pixels? and y is still amplitude?

• Well, systems that work on 2D data are 2D systems, therefore there's no single $x$-axis, but the impulse response is a function of two coordinates $x_1,x_2$, just like intensity of pixels is a function of two coordinates in an image. Nov 14 '20 at 16:15
• yeah, I got the point! Thank you
– Mark
Nov 14 '20 at 16:24

You imagine several representations. A first one is the standard image representation: flat in the $$(x_1,x_2)$$ and a color scale of the intensity value (left plot). A second one is a 3D vision as a topographical or elevation map, with homogeneous color (middle plot). The visualization might be a bit more complicated. A third one combine the two preceding aspects (right plot). The latter helps understanding some mathematical morphology operators (like the watershed).

L = membrane(1,100);
f = figure;
ax = axes;
subplot(1,3,1)
imagesc(L);
subplot(1,3,3)
s = surface(L);
s.EdgeColor = 'none';
view(3);axis tight
subplot(1,3,2)
s = surface(L);
s.EdgeColor = 'none';
s.FaceColor = 'b';
view(3);axis tight

• yeah, it really helps. so it depends on the way of represing it
– Mark
Nov 14 '20 at 16:42
• It really depends on the nature of the data, and what you'd like to visualize. Other types of visualization are still possible Nov 14 '20 at 16:44