how to assign imagesc to figure 1

In this part you will write a program to perform a simple analysis of a random image. First, generate a random array of size 40x40 with each entry being either 0 or 1 and display it in Figure 1. Next, write a program segment that searches for single entries (value = 1) that have no neighbors (all 8 neighbors have values of 0) and records them in another array of the same size. Display the results visually in Figure 2. An example run is shown below: content of random matrix (left) and entries with no neighbors (right).

Please help me with underlined part

n = 40;
m = 40;
A1 = rand(n,m) < 0.5;
imagesc(A1);

Answer 1

[Walter Roberson]

fig1 = figure(1);
ax1 = axes('Parent', fig1);

imagesc(A1, 'Parent', ax1);

If there is an existing axes you want to draw into, then use its handle in the 'Parent' option of imagesc

[Walter Roberson]

I suggest you read about conv2() with regards to finding isolated pixels.

Answer 2

[Image Analyst]

Following Walter's code call figure(2) to create a figure numbered "2", then use bwmorph() with the 'clean' option to remove single pixels. Use xor() with the original binary image to get an image with only the single, isolated pixels.

fig1 = 
A1 = 
imshow(A1)
output = bwmorph(
output = xor(
fig2 = figure(2);
imshow(output)

Complete this since it seems to be your homework and they probably don't want you to turn in our solution as your own. It should be trivial now that we've done most of it for you.

[Walter Roberson]

No, no -- bwareafilt() with an area range of 1 to 1 !

[Walter Roberson]

why are we using bw when we dont want it black and white

Your array is defined as consisting only of 0 and 1. That is black (0) and white (1)

[Walter Roberson]

output = xor(1,1);

You take the scalar value 1, and you xor it with the scalar value 1. The result is going to be a scalar 0.

Note: Image Analyst recommended bwmoprh(), not bwmorph3(). And he recommended xor'ing the result of bwmorph() with the original image.

[Image Analyst]

I'm confused. My guess is that @Eiman Rana deleted some comments between Walter's.

Also, there are multiple ways to achieve the same result, some of which depend on the data type of the image. I gave one way but there are others. Looks like bwareafilt() could do it in one line of code instead of 2 with bwmorph() and xor().

[Walter Roberson]

output = conv2(A1,  [1 1 1;1 -1 1; 1 1 1], 'same') == -1;