How i can convert any type of image to ASCII values and store these values into one-dimensional array?

Hello matlab community.

I want to convert any type of image ( color or gray scale image ) to ASCII values and store these values into one-dimensional array for ex., if read my image as: A = imread(filename);, i want the final results(ASCII values ) appears in one-dimensional array(ex: B=[150,146,....), so that later on, I can reconstruct the original image.

Answer 1

[Image Analyst]

I don't understand why you want to do that. If you want the original image later, just make a copy of A before you alter it.

You can make a 1-D list B from A like this

B = A(:)

Not sure what you're going to do with that though if A is a color image. If you have a gray scale image it's just column 1, then column 2, then column 3, etc. until the last column of the image, but I'm not sure why you'd want this.

If you want to make an ASCII text file with the location and color value in a CSV file, see the attached demo.

[Radhwan Jawad]

consider the final result is :

R G B X Y

200 125 140 1 1

143 145 142 1 2

139 120 255 1 3

i want the output is one-dimensional array for ex. :

output=[200, 125,140,143,145,142,139,120,255]

and i will make some operations on (output) array, then i need to reconstruct the image, how can i doning this?

[Radhwan Jawad]

And thank you very much for the answer

[Walter Roberson]

i want the output is one-dimensional array for ex. :

output=[200, 125,140,143,145,142,139,120,255]

In that case

output = reshape( permute(YourImage, [3 1 2]), 1, []);

reconstructed = permute(reshape(output, SIZE_OF_IMAGE(3), SIZE_OF_IMAGE(1), SIZE_OF_IMAGE(2)), [2 3 1]);

[Radhwan Jawad]

Thank you so much.

I trided your code(final_output = reshape( permute(rgbImage, [3 1 2]), 1, []);) on (RGBXY_to_CSV.m) file and your code works very well, but i not understand what did you mean (SIZE_OF_IMAGE()) in secode line of your code, please help me.

[Image Analyst]

SIZE_OF_IMAGE = size(YourImage);

where you of course replace "YourImage" with the actual name you used for your image variable.

Or you could do this:

[rows, columns, numberOfColorChannels] = size(YourImage)
output = reshape( permute(YourImage, [3 1 2]), 1, []);
reconstructed = permute(reshape(output, numberOfColorChannels, rows, columns), [2 3 1]);

[Radhwan Jawad]

Thank you so much..

Answer 2

[Walter Roberson]

N = ndims(YourImage);
indices = cell(1,N);
idx = find(YourImage);
[indices{:}] = ind2sub(size(YourImage), idx);
values = YourImage(idx);

T = [values(:), indices{:}];

Now T is an array with multiple columns. The first column is a pixel value; the remaining columns are indices -- row, column at the very least; if the 4th column is present it is channel number.

This code does not record the location of any pixels that are zero.

[Radhwan Jawad]

And thank you very much for the answer

but, how i can reconstruct the original image?

[Walter Roberson]

reconstructed = accumarray(T(:,2:end), T(:,1));

This will not work exactly as desired if it happens that there are rows or columns along the right edge or top edge that are all black: for that case you would need to know what the original size of the image is.