how to plot a graph in matlab ?

Let given array A with 512*512 elements are (here one sample of 4*4 is given)

5    4     6     8
9   11    30    34
12  13    13    14

how to plot a graph where x-axis contains the values of individual elements of 512*256 matrix and y-axis contain the frequency of the difference value. The frequency of difference value can be obtained as

   1   -2
  -2   -4
  -1   -1

here the difference value can be found by subtracting the 2 consecutive elements in a row.Now suppose the original matrix A is 512*512 elements, then the difference matrix will be 512 * 256 elements.

Kindly suggest

[KSSV]

Do you have any pictorial example?

[Walter Roberson]

A(:, 1:2:end) - A(:,2:2:end) would give you a 512 x 256 array.

You have not indicated why the output should only be 256 rows.

[aditya sahu]

I m really sorry ...the output should be 512 * 256 only..

[aditya sahu]

ok..by doing A(:, 1:2:end) - A(:,2:2:end) gives 512 x 256 array. Then how can i plot a graph with x-axis as its elements of 512*256 array and y-axis as its frequency of occurence.

[aditya sahu]

@ KSSV SIR, I HAVE ATTACHED A SAMPLE FILE THE PLOT SHOULD COME OR LOOKS LOOKS LIKE THIS.

Answer 1

[Walter Roberson]

   B = A(:, 1:2:end) - A(:,2:2:end);
   histogram(B, 'binmethods', 'integers')

Note: this might require a fairly recent MATLAB (I do not recall ever having seen that option before.)

[aditya sahu]

Dear sir, when i use, histogram(B, 'binmethods', 'integers') The error is comming as Undefined function 'histogram' for input arguments of type 'uint8'. Kindy tell what has to be done now.?

[Walter Roberson]

Histogram was introduced in R2014b, and has always supported uint8. On the other hand, subtracting uint8 data can never give you negative values, so even if you were using a new enough version then for uint8 data you should use

     B = double(A(:, 1:2:end)) - double(A(:,2:2:end));

For versions older than R2014b:

bins = min(B(:)) : max(B(:));
hist(B, bins);

[aditya sahu]

Dear @ Walter Roberson thank for your help. but the below code is working.But only problem is in x-axis i want to limit it to (-40 to 40) ..can you help to sort this issue.

z=double(int32(A(:, 1:2:end))) - double(int32((A(:,2:2:end))))
  [a,b]=hist(z,unique(z));
  out=[b' sum((a),2)]
  x=out(:,1);
  y=out(:,2);
  plot(x,y,'--bo')

[Walter Roberson]

z = double(A(:, 1:2:end)) - double(A(:,2:2:end));

bins = -40:40;
N = histc(z, bins);

x = bins;
y = N;
plot(x, y, '--bo')

[aditya sahu]

No sir your code is giving incorrect plot ..kindly tell in my code how can i restrict the x-axis range to -40 to 40.The code is

z=double(int32(b(:, 1:2:end))) - double(int32((b(:,2:2:end))));

    [a,c]=hist(z,unique(z));
    out=[c' sum((a),2)];
    x=out(:,1);
    y=out(:,2);
    plot(x,y, '--bo')

[aditya sahu]

ya..i got it...

plot(x(x>-40 & x<40), y(x>-40 & x<40),'or')

..is it right..

[Walter Roberson]

why are you converting to int32 before converting to double??

z = double(A(:, 1:2:end)) - double(A(:,2:2:end));

bins = -40:40;
N = histc(z(:), bins);

x = bins;
y = N;
plot(x, y, '--bo')

[aditya sahu]

Thank you dear @ Walter Roberson.. Actually not required,but i have taken two different matrices where one is in uint8 and other is double. thats i have brrought back both to unit32..Thats all only.

[Walter Roberson]

It is not a problem to use double() on an array that is already double(), so leave out the conversion to int32()