Please help me out in rectifying the errors in implementing Content Based Image Retrieval System.

I have 56 images in a database.i want to display and coompare best five images with the query image using eucledian distance.the problem which i get is that the result doesnot show best five images instead its showing only one image which is the query image.my program is:

PROGRAM NO 1:

clc;

G=imread('n252.tif');

H = adapthisteq(G);

[rows cols]=size(G);

[c1,s1]=wavedec2(H,1,'db1');

%disp(c1);

X=c1;

figure,imshow(G);

%figure,imshow(H);

fileFolder=fullfile(matlabroot,'toolbox','images','imphoto');

dirOutput=dir(fullfile(fileFolder,'*.tif'));

fileNames={dirOutput.name}

n=numel(fileNames)

g=zeros(1,n);

for k = 1 : n

fileNames1=strcat('fullfile(fileFolder)',fileNames(k))

I = imread(fileNames{k});

J = adapthisteq(I);

J = imresize(J, [rows cols]);

[c2,s2]=wavedec2(J,1,'db1');

%disp(c2);

Y=c2;

%[rows1 cols1]=size(J);

E_distance = sqrt(sum((X-Y).^2));

g(1,k)=E_distance;

if g(1,k)==0

    w=k;

end

%figure,imshow(I);

%figure,imshow(J);

end

disp(g);

%II=imread(fileNames{w});

%figure, imshow(II);

[sorted,IX] = sort(g);

bestFiveImages = IX(1:5);

for I = 1:length(bestFiveImages)

 figure;imshow(imread(fileNames{bestFiveImages(I)}));

end

the error came as follows:

??? Error using ==> images\private\checkinput>check_attributes

Function ADAPTHISTEQ expected its first input argument, I,

to be two-dimensional.

Error in ==> images\private\checkinput at 37

check_attributes(A, attributes, function_name, variable_name, ...

Error in ==> adapthisteq>parseInputs at 411

checkinput(I, {'uint8' 'uint16' 'double'}, ...

Error in ==> adapthisteq at 155

[I, selectedRange, fullRange, numTiles, dimTile, clipLimit, numBins, ...

Error in ==> new at 18

J = adapthisteq(I,'clipLimit',0.02,'Distribution','rayleigh');

please help me out in rectifing these errors..

actualy i have done the same program but with different database.in that case the result came out but when i have changed the database the above problem persists..

the same program which is done by using different database is as follows(here the result came):

PROGRAM NO 2:

clc;

G=imread('lion1.jpg');

G1=rgb2gray(G);

H = adapthisteq(G1);

[rows cols]=size(G1);

[c1,s1]=wavedec2(H,1,'db1');

%disp(c1);

X=c1;

figure,imshow(G);

%figure,imshow(H);

fileFolder=fullfile(matlabroot,'toolbox','images','imda');

dirOutput=dir(fullfile(fileFolder,'*.jpg'));

fileNames={dirOutput.name}

n=numel(fileNames)

g=zeros(1,n);

for k = 1 : n

fileNames1=strcat('fullfile(fileFolder)',fileNames(k))

I = imread(fileNames{k});

I1=rgb2gray(I);

J = adapthisteq(I1);

J = imresize(J, [rows cols]);

[c2,s2]=wavedec2(J,1,'db1');

%disp(c2);

Y=c2;

%[rows1 cols1]=size(J);

E_distance = sqrt(sum((X-Y).^2));

g(1,k)=E_distance;

if g(1,k)==0

    w=k;

end

%figure,imshow(I);

%figure,imshow(J);

end

disp(g);

%II=imread(fileNames{w});

%figure, imshow(II);

[sorted,IX] = sort(g);

bestFiveImages = IX(1:5);

for I = 1:length(bestFiveImages)

 figure;imshow(imread(fileNames{bestFiveImages(I)}));

end

here in program no 2 no such error is coming.so i want to know what is the mistake in program no 1.the result is coming in program no 2 but in program no 1 no result is coming instead the errors are coming.though the two programs are same only the difference is that i have taken different databases in these two above mentioned programs..Thanks

[Rima]

please help me @image analyst ,@walter roberson

[Iman Ansari]

The error says your image I is not 2 dimensional (grayscale) image. In second code you used rgb2gray before adapthisteq function:

I1=rgb2gray(I);
J = adapthisteq(I1);

[Rima]

yeah i know the error but how to rectify them?please help and i used rgb2gray in 2nd code because some of the images are coloured but i want gray image.thanks

Answer 1

[Image Analyst]

adapthisteq() doesn't work with color images.

Anyway, there is no CBIR that I can see in your code. You're just comparing the intensities of the wavelet images. I could be wrong because I'm not a wavelet guy, but I don't think that will allow you to do things like "show me all pictures with a beach in them" or "show me all photos with a dog in them", etc.

[Rima]

i have used this rgb2gray in the program no 1 but still the error comes..now the error shows:

??? Error using ==> rgb2gray>parse_inputs MAP must be a m x 3 array.

 Error in ==> rgb2gray at 35 X = parse_inputs(varargin{:});

 Error in ==> new3 at 3 G1=rgb2gray(G); >>

[Iman Ansari]

if size(I,3)==3
  I=rgb2gray(I);
end

[Image Analyst]

Evidently instead of just passing back the input image if it's already grayscale (like one might intuitively think), rgb2gray errors out. So to avoid that check how many color channels it has, and call it only if it's color, like Iman showed you.

[Rima]

thanks @Iman and @image analyst...yes it worked..actually there were coloured image in my database.