Converting image to bits and vice versa

Hi all. I am designing convolutional encoder by using Matlab. To see the result clearly, I will use some pictures. I need to convert the picture pixels into binary code and after I encode it, I need to convert these binary codes into pixels and plot it. Can you help me for convertion part?

[Rik]

If you are planning to store them in 8 bit bytes you can use the uint8 data type and use the normal image displaying tools.

Answer 1

[Walter Roberson]

bits = reshape((dec2bin(typecast(YourImage, 'uint8'), 8) - '0').', 1, []);

orig_class = class(YourImage);
orig_size = size(YourImage);

reconstructed = reshape(typecast(uint8(bin2dec(char(reshape(bits, 8, [])+'0').')), orig_class), orig_size);

In the special case that you are willing to restrict your images to be uint8, you can simplify these steps.

Unless you are willing to restrict yourself to one fixed image size, you need to convey the original image size to the remote end somehow.

[Andy Paulo Ureña]

Hello Walter, can you help me with the error that im having with your code ? I just put the variable that store the image that i want to convert. Thanks

Error using typecast

The first input argument must be a vector.

Error in imageprocessing (line 37)

bits = reshape((dec2bin(typecast(img, 'uint8'), 8) - '0').', 1, []);

[Walter Roberson]

YourImage = imread('peppers.png');
imshow(YourImage); title('original');

bits = reshape((dec2bin(typecast(YourImage(:), 'uint8'), 8) - '0').', 1, []);

orig_class = class(YourImage);
orig_size = size(YourImage);

reconstructed = reshape(typecast(uint8(bin2dec(char(reshape(bits, 8, [])+'0').')), orig_class), orig_size);
imshow(reconstructed); title('reconstructed')

max(abs(double(YourImage(:)) - double(reconstructed(:))))

ans = 0