plot histogram without using matlab hist() function

Hi all

There is a function called hist(), but if I want to plot a graph without using hist() command, how to plot a graph? I means using command to do the job same as hist().

Thank you.

[Aditi Shetty]

    A=imread('<your image path>');
    a=rgb2gray(A);
    subplot(3,1,1);
    imshow(a);
    title('original image');

    [r,c]=size(a);
    z=zeros(1,256);
    for i=1:r
        for j=1:c
            b=a(i,j);
            z(b+1)=z(b+1)+1;
        end
    end

    N=sum(z);
    p=zeros(1,256);
    s=zeros(1,256);
    c=zeros(1,256);
    r=zeros(1,256);

    for k=1:256
        p(k)=z(k)/N;
        if k==1
            c(k)=p(k);
            s(k)=c(k)*255;
            r(k)=floor(s(k));
        else
            c(k)=c(k-1)+p(k);
            s(k)=c(k)*255;
            r(k)=floor(s(k));
        end
    end
    subplot(3,1,3)
    stem(r,z)
    title('histogram')

[Ritvik Ramesh Palvankar]

Hey Adil,

Can you explain me the code after "for k=1:256". I did not understand the part where you took s(k)=c(k)*255

Also, why is zero matrix used and why is c(k-1) added to p(k)

Answer 1

[Naz]

It depends on how you want to bin your data and what kind of data you have.

The idea of hist function is similar to stacks of coins: Let's say you have a bag of coins and after you separate them by values and stack on one another, you will get different heights. The hist function does the same.

data=[1 3 5 7 4 8 0 1 3];

If you want to distribute your data into 10 bins, you would create a new array of size 10:

histArray=zeros(1,10);  % prealocate
x=0:1:9;
then, you would run a forloop to count how many times you encounter in a particular value:

for n=1:length(data)
    histArray(1,data(n)+1)=histArray(1,data(n)+1)+1; % every time you meet the particular value, you add 1 into to corresponding bin
end
bar(histArray)

There could be another case. Let's say you want to distribute your data into 5 bins only (0 and 1 go in the 1st bin, 2 and 3 go in the 2nd...). In this case, you would run the same loop but with more complicated conditioning:

    histArray=zeros(1,5);  %prealocate
    x=0:2:8  % create bin numbers for plot
    for n=1:length(data)
        histArray(1,floor(data(n)/2)+1)=histArray(1,floor(data(n)/2)+1)+1;
        % every time you meet the particular value
        %you add 1 into to corresponding bin
    end
    bar(x,histArray)

You can design your own hist function to fit you needs

[ws]

thanks, this is what i'm looking for.. :)

[Anna Kasdan]

Hi! I am trying to do this with data that includes negative numbers, and I get the error 'Subscript indices must either be real positive integers or logicals.' Any thoughts? The data I am working with is:

n = 200; a = 0; b = 2; data = b.* randn(1, n) + a;

[Sahitya Bonumaddi]

i'm stuck here too

Answer 2

[Adrian Lukasik]

hello I have another problem. The same thing, but in 3D: There is a function called hist3(), but if I want to plot a graph without using hist3() command, how to plot a graph? I means using command to do the job same as hist3(). I have matrix b, 2x1000, and the task about making histogram plot without hist3() function.

Thank you.

[Walter Roberson]

You should open a new Question for that.

Once you have the counts for each grid-point, use bar3()

Answer 3

[Saad Alzubaidi]

clear all,clc;
x=imread(imgetfile);
imtool(x)
for r=1:3
    v=x(:,:,r); 
    h(1:255,3)=0;
    for i=1:size(x,1)
        for j=1:size(x,2)
            h(v(i,j)+1,r)=h(v(i,j)+1,r)+1;
        end
    end
end
figure,plot(h);

[Saad Alzubaidi]

This is the code you requested, for the color image.