bounjours a tous , alors j'ai commence de trvaille avec matlab j'ai esseye de faire le produit de convolusion entre une image et un filtre gaussian avec la fonction conv2 mais je suis bloqué a cause de ca voila l'erreur:N-D arrays are not supported.

sigma1 = 1.9; 
sigma2 =2.8;
img=imread('image.jpg');
hsize = [3,3];
h1 = fspecial('gaussian', hsize, sigma1);
h2 = fspecial('gaussian', hsize, sigma2);
% [a,b]=size(img);
% m1=max([a+3-1,a,3]);
% n1=max([b+3-1,b]);
% im1=cell(m1,n1);
im1=conv2(img,h1);
im2=conv2(img,h2);

Answer 1

[Rik]

Next time, use the {}Code button to format your code.

You did img=imread('image.jpg');, which results in a 3D matrix (column,row,color). You need to reduce that to two dimension, e.g. with img=imread('image.jpg');img=mean(img,3);.

[Rik]

Si mon répons t'as aidé, merci de l'accepter, si non, tu peut écrire un répons ici, et on essaiera a t'aider.

[amal abbbes]

thank you for your help Rik, now it is ok

[amal abbbes]

now i'have other probleme in my code i want to create a pyramide de gaussian But before I want to cut a part of the image with the function Imcrop then I want applied on this image a product of convolusion with the filter Gaussian then I want to stored the result of difference of two image of in a matrix [1 3] that contains in each case E of the sub matrix and here is the problem in this level. error:Assignment has more non-singleton rhs dimensions than non-singleton subscripts Error in mat (line 59) temp_D(k,i,j)=L2-L1;

    img=imread('img1.jpg');
[a,b,p]=size(img);
if p==3
  img=rgb2gray(img);
end   
IMG= imcrop(img,[1 1 660 578]);
[row colum]=size(img);
c=size(img,3);
img=im2double(img);
origin=img;
toc
%%Scale-Space Extrema Detection
tic
% original sigma and the number of actave can be modified. the larger
% sigma0, the more quickly-smooth images
sigma0=sqrt(2);
octave=3;%6*sigma*k^(octave*level)<=min(m,n)/(2^(octave-2))
level=3;
D=cell(1,octave);
for i=1:octave
D(i)=mat2cell(zeros(row*2^(2-i)+2,colum*2^(2-i)+2,level),row*2^(2-i)+2,colum*2^(2-i)+2,level);
end
% first image in first octave is created by interpolating the original one.
temp_img=kron(img,ones(2,2));
temp_img=padarray(temp_img,[1,1],'replicate');
figure(2)
subplot(1,2,1);
imshow(origin)
%create the DoG pyramid.
for i=1:octave
    temp_D=D{i};
    for k=1:3
    for j=1:level
        scale=sigma0*sqrt(2)^(1/level)^((i-1)*level+j);
        p=(level)*(i-1);
        figure(1);
        subplot(octave,level,p+j);
        f=fspecial('gaussian',[7,7],scale);
        L1=temp_img;
        if(i==1&&j==1)
        L2=conv2(temp_img,f,'same');
        L2=conv2(L2,f','same');
        temp_D(k,i,j)=L2-L1;
        imshow(uint8(255 * mat2gray(temp_D(:,:,j))));
        L1=L2;
        else
        L2=conv2(temp_img,f,'same');
        L2=conv2(L2,f','same');
        temp_D(:,:,j)=L2-L1;
        L1=L2;
        if(j==level)
            temp_img=L1(2:end-1,2:end-1);
        end
        imshow(uint8(255 * mat2gray(temp_D(:,:,j))));
        end
    end
    D{i}=temp_D;
    temp_img=temp_img(1:2:end,1:2:end);
    temp_img=padarray(temp_img,[1,1],'both','replicate');
    end
end
]
end
end
  end

[Rik]

It looks like L2 and L1 are 2D matrices, while your assignment only works for a scalar. A bit later in your code you're using temp_D(:,:,j)=L2-L1; which should work here as well, although I don't know if that is what you mean.

[amal abbbes]

Here I want to created a single-line matrix and three columns with each box containing another matrix I want to stores the image in this sub matrix.so How I can solve the problem please

Answer 2

[Image Analyst]

You need to convert from color to gray scale and then cast the image to single or double before sending in to conv2().

Try this:

clc;
fontSize = 15;
% Read in image.
rgbImage = imread('peppers.png');
% Display the image.
subplot(2, 3, 1);
imshow(rgbImage, []);
title('Original Image', 'FontSize', fontSize, 'Interpreter', 'None');
axis on;

% Get the dimensions of the image.  
% numberOfColorChannels should be = 1 for a gray scale image, and 3 for an RGB color image.
[rows, columns, numberOfColorChannels] = size(rgbImage);
% Convert to gray scale.
if numberOfColorChannels > 1
  % It's not really gray scale like we expected - it's color.
  % Use weighted sum of ALL channels to create a gray scale image.
%   grayImage = rgb2gray(rgbImage); 
  % ALTERNATE METHOD: Convert it to gray scale by taking only the green channel,
  % which in a typical snapshot will be the least noisy channel.
  grayImage = rgbImage(:, :, 1); % Take red channel.
else
  grayImage = rgbImage; % It's already gray scale.
end
% Display the image.
subplot(2, 3, 2);
imshow(grayImage, []);
title('Gray Scale Image', 'FontSize', fontSize, 'Interpreter', 'None');
axis on;

% Get the blur kernels.
sigma1 = 1.9; 
sigma2 = 2.8;
windowSize = [3,3];
h1 = fspecial('gaussian', windowSize, sigma1);
h2 = fspecial('gaussian', windowSize, sigma2);
% Display the image.
subplot(2, 3, 3);
imshow(h1, []);
title('Blur Kernel 1 Image', 'FontSize', fontSize, 'Interpreter', 'None');
axis on;
% Display the image.
subplot(2, 3, 4);
imshow(h2, []);
title('Blur Kernel 2 Image', 'FontSize', fontSize, 'Interpreter', 'None');
axis on;
drawnow; % Force screen repaint right now.

% Blur the image.
im1 = conv2(double(grayImage), h1);
im2 = conv2(double(grayImage), h2);

% Display the image.
subplot(2, 3, 5);
imshow(im1, []);
title('Blurred Image 1', 'FontSize', fontSize, 'Interpreter', 'None');
axis on;
% Display the image.
subplot(2, 3, 6);
imshow(im2, []);
title('Blurred Image 2', 'FontSize', fontSize, 'Interpreter', 'None');
axis on;
% Enlarge figure to full screen.
set(gcf, 'Units', 'Normalized', 'Outerposition', [0, 0.05, 1, 0.95]);

[amal abbbes]

thanks you for your help ,it run :) now i'have other probleme in my code i want to create a pyramide de gaussian But before I want to cut a part of the image with the function Imcrop then I want applied on this image a product of convolusion with the filter Gaussian then I want to stored the result of difference of two image of in a matrix [1 3] that contains in each case E of the sub matrix and here is the problem in this level. error:Assignment has more non-singleton rhs dimensions than non-singleton subscripts

Error in mat (line 59) temp_D(k,i,j)=L2-L1;

    img=imread('img1.jpg');
[a,b,p]=size(img);
if p==3
  img=rgb2gray(img);
end   
IMG= imcrop(img,[1 1 660 578]);
[row colum]=size(img);
c=size(img,3);
img=im2double(img);
origin=img;
toc
%%Scale-Space Extrema Detection
tic
% original sigma and the number of actave can be modified. the larger
% sigma0, the more quickly-smooth images
sigma0=sqrt(2);
octave=3;%6*sigma*k^(octave*level)<=min(m,n)/(2^(octave-2))
level=3;
D=cell(1,octave);
for i=1:octave
D(i)=mat2cell(zeros(row*2^(2-i)+2,colum*2^(2-i)+2,level),row*2^(2-i)+2,colum*2^(2-i)+2,level);
end
% first image in first octave is created by interpolating the original one.
temp_img=kron(img,ones(2,2));
temp_img=padarray(temp_img,[1,1],'replicate');
figure(2)
subplot(1,2,1);
imshow(origin)
%create the DoG pyramid.
for i=1:octave
    temp_D=D{i};
    for k=1:3
    for j=1:level
        scale=sigma0*sqrt(2)^(1/level)^((i-1)*level+j);
        p=(level)*(i-1);
        figure(1);
        subplot(octave,level,p+j);
        f=fspecial('gaussian',[7,7],scale);
        L1=temp_img;
        if(i==1&&j==1)
        L2=conv2(temp_img,f,'same');
        L2=conv2(L2,f','same');

        temp_D(k,i,j)=L2-L1;
        imshow(uint8(255 * mat2gray(temp_D(:,:,j))));
        L1=L2;
        else
        L2=conv2(temp_img,f,'same');
        L2=conv2(L2,f','same');
        temp_D(:,:,j)=L2-L1;
        L1=L2;
        if(j==level)
            temp_img=L1(2:end-1,2:end-1);
        end
        imshow(uint8(255 * mat2gray(temp_D(:,:,j))));
        end
    end
    D{i}=temp_D;
    temp_img=temp_img(1:2:end,1:2:end);
    temp_img=padarray(temp_img,[1,1],'both','replicate');
    end
end
]

end

end

[Image Analyst]

You forgot to attach img1.jpg. Also, there is an extra right bracket:

end
]

end

that I'm sure is causing problems. And in

temp_D(k,i,j)=L2-L1;

L2 and L1 are 2-D images and you're trying to stuff the difference of two images, which is itself another image, into a single element of a 3-D array. Why are you trying to do that? Maybe you should make temp_D a cell array if it's supposed to hold different sized arrays.

[amal abbbes]

it is as if Submatrix Nested in one each box of the matrice.I wants to create a matrix m (1, 3) which contains a single row and three columns and in each box of the matrix m, m (i, J) I want to create another m2 matrix of size (3.1) and then in Ach cell of the M2 matrix ,I want to store the image :pyramide de gaussaian how can i do it please