How to represent a 2D matrix as a grayscale image

i have a 2D matrix and i would like to visualize it as a grayscale image. i used the fuction mat2gray(x) and then imshow. However, i had got black and whites squares as shown in the first image attached below and i want to merge those squares and to get the image as the second one.

[DGM]

If that's your input and desired output, what exactly do you mean by "merge"? Do you intend to simply sort the pixels?

[salim sak]

i want to get the output as the following image:

[salim sak]

you can ignore the two white-colored lines and focus on other things

[AndresVar]

@salim sak it seemed like you just wanted to sort the pixels, but this last image looks like there is a repeating pattern and white line. Is there a function or is it just randomly create gradient patterns?

[DGM]

I can ignore the white lines, but that still doesn't help clarify the conceptual connection between the first image and either of the second or third images. How are the images related?

Wait a minute....

Where did the data come from? Is this an indexed image? Is there an associated colormap or an image copy from which the map can be derived?

[DGM]

The third image is a plot. The caption reads:

The classification accuracy of five classifiers per three combinations of electrode (C1, C2, C3)...

So is that related to what the first image is?

[Rik]

You could also simply edit your previous question instead of deleting it and posting a new one.

[Image Analyst]

Your first, second, and third posts, each showing an image, are all different images. I'm not sure if or why they're supposed to be different, but they are.

A matrix already IS an image. You can display any 2-D image with

imshow(M, []);

Answer 1

[Walter Roberson]

Look at your top left square of your "black and white" image. It is clearly not the same shade as the square to its right, and neither of them is the same shade as the square under the top left square. Therefore you do not have a "black and white" image, you have a grayscale image in which adjacent locations happen to have notable differences.

If we create some data with notable differences between adjacent locations:

data = randn(8,8);
scaled_data = mat2gray(data);
figure(); imshow(scaled_data)

Now for the purposes of demonstration we sort the data:

figure(); imshow(sort(sort(scaled_data,1,'ascend'),2,'descend'))

and we see something that looks more like what you are asking for. Both cases are using imshow(), so it is not imshow() that is making the difference: the difference is that the data happens to be arranged differently so that there are not high contrast between neighbours.

In short, the problem is not with imshow(): the problem is with your data happening to have high contrast between adjacent locations.

Maybe you want interpolation to reduce the contrast?

figure(); pcolor(scaled_data); colormap(gray)