error in plotting a bar graph

this program implements a fuzzy color histogram,the fuzzy system a takes 3 inputs.rgb2lab is already defined.the output graph,i.e the histogram plot has 15 bins

 z=imread('Lenna.png');

x=readfis('a.fis');
R=z(:,:,1);
G=z(:,:,2);
B=z(:,:,3);
%x1=0:1:15;
[L,a,b]=RGB2Lab(R,G,B);

%L1=cat(3,L,a,b);
%imshow(L1)
[m,n,o]=size(s);
q=[];
%L=zf(:,1);
Ld=double(L);
%L1d=double(L1);
%a=zf(:,2);
ad=double(a);
%b=zf(:,3);
bd=double(b);
for i=1:m
    for j=1:n

              q(i,j)=evalfis([Ld(i,j) ad(i,j) bd(i,j)],x);

      end
  end

bar(q(i,j))

i need to plot the output 'q' as a bar graph.and i need the x-range to be in 0-15. but there is an error: rror using ==> bar The length of X must match the number of rows of Y. can anyone help with this?

[piyali banerjee]

when running this code I am getting q[] values same for all the rows and columns as .5 for any jpg images.Can anybody tell me which portion of the code should I modify to get the correct result aswellas correct bar chart?

Answer 1

[Walter Roberson]

You have a nested for loop over i and j. After the end of that loop, you ask to bar(q(i,j)) . That is asking to create a bar plot of the single value stored at q(i,j) based upon the values of i and j after the "for" loops. MATLAB leaves the loop variable in a "for" loop set to the last valid value, so this will be q(m,n) that you are asking to create a bar plot of.

Considering that q is 2 dimensional, it is not at all clear what you are wanting to create a bar plot of. Especially as, in past threads, you have indicated that what evalfis() on your particular FIS returns is already the histogram, rather than a histogram bin number.

If I were to guess, I would say that what you want is

hist(q(:))

[mideas]

The FIS returns the histogram values. N i want to plot those values as a bar graph.

And hist(q(:)) gives error. Thanx anyways.

[Walter Roberson]

Is the error you get for hist(q(:)) the same as you reported earlier, "The length of X must match the number of rows of Y" ? If so then you either have a corrupted version of bar(), or else (more likely) you have accidentally named one of your own .m files with the name of a routine that bar() calls upon for its computations.

Note: considering the range of values you want, I should have said to use

hist(q(:), 16)

That would only change the visible output though, and will not in itself fix any problem with bar()

I suggest that at the command line, you command

dbstop in bar

and then run the hist(q(:),16) . When it stops in bar.m, step through line by line to see which routines are being called, and make sure that you do not have any .m routines of your own with the same name.

[mideas]

this is the error that is shown( for hist(q(:)) and hist(q(:),16) ):-

Error using ==> mtimes

Integers can only be combined with integers of the same class, or scalar doubles.

Error in ==> hist at 63

xx = miny + binwidth*(0:x);

And i dont have any .m routines with the same name.

[Walter Roberson]

Hmmm... try hist(double(q(:)),16)

[mideas]

i think its working now.. thanx a lot!

[piyali banerjee]

I am getting all the q(m,n) values as .5 taking any images as consideration.Can anyone solve the problem?