To convert 3D color image to 2D .

I am trying to denoise color images using mean, median, adaptive filters. But there is this error: ?? Undefined function or method 'conv2' for input arguments of type 'double' and attributes 'full 3d real'.

I need the output images in color scale. So 'rgb2gray' is ruled out. Any better options other than 'rgb2ind' and 'extracting color channels'to debug this code.

Answer 1

[Image Analyst]

Just cast to double

% Extract the individual red, green, and blue color channels.
redChannel = rgbImage(:, :, 1);
greenChannel = rgbImage(:, :, 2);
blueChannel = rgbImage(:, :, 3);
% Blur them
kernel = ones(7)/49;
blurredRedChannel = conv2(double(redChannel), kernel, 'same');
blurredGreenChannel = conv2(double(greenChannel), kernel, 'same');
blurredBlueChannel = conv2(double(blueChannel), kernel, 'same');

Answer 2

[Walter Roberson]

If you need to process a 3D array in your convolution call then you need to call convn() instead of conv2()

Image Analyst's code processes only a 2D slice of the 3D array, and so does not handle the case that you need to consider the color components linearly in your convolution.

If you want to process each 2D slice with the same 2D filter then you can call conv2 in a loop, or you could call convn with a convolution array that happens to be 2D.

[Chithra Shruthi]

I have not used conv2 in my code. The conv2 error is because I have used 'filter2' which uses conv2.

I need to know how I could change the dimensions of the color image so that I could use 'filter2' , 'medfilt2', 'wiener2' on the color image.

Here is my code:

clc;
clear all;
close all;

I=imread('Color image.jpg');
subplot(2,3,1), imshow(I);
title('Original Image');

Jm=imnoise(I,'salt & pepper', 0.02);
subplot(2,3,2), imshow(Jm);
title('Image corrupted by Salt and Pepper noise');
J= im2uint8(Jm);

K=filter2(fspecial('average', 3), J)/ 255;
subplot(2,3,3), imshow(K);
title('Average filter');

L=medfilt2(J,[3 3]);
subplot(2,3,4), imshow(L); 
title('Median filter');

M=wiener2(J, [5 5],0.01);
subplot(2,3,5), imshow(M);
title('Adaptive filter');

B= fspecial('gaussian', [5 5], 2);
fin= imfilter(J,B, 'same');
subplot(2,3,6), imshow(fin);

And this is the error:

??? Undefined function or method 'conv2' for input arguments of type 'double' and
attributes 'full 3d real'.

Error in ==> filter2 at 73
        y = conv2(hcol, hrow, x, shape);

Error in ==> colorsnp at 18
K=filter2(fspecial('average', 3), J)/ 255;

[Walter Roberson]

http://www.mathworks.com/matlabcentral/answers/17529-what-is-the-difference-between-conv2-filter2-and-imfilter

[Image Analyst]

Then use conv2(), like I showed you. You can also use it on the intact RGB image and it will do it a color channel at a time.

[Walter Roberson]

You cannot change the dimensions of your color image to use filter2 and so on, at least not and get useful results.

You can extract individual color planes from the image and do the processing on those, the way that Image Analyst showed.

But if you insist... then reshape the color image to rows by (columns times 3) to get a 2D array that you can filter, and eventually reshape the result back to 3D. You will have some odd edge effects. But that is not a problem. You have not defined the interactions that you want between the color planes so weird edge interactions are not against any requirement you have given so far.