The first thing where you've gone wrong is to create all these C* images. If you're numbering variables, you're doing something wrong. Matlab already has a very effective way of storing things with a numbering index, it's called a matrix or cell array. So:
Cbit{1} = bitget(C, 1); figure, ...
Cbit{2} = ...
%etc
And, since you're now using a cell array, you can avoid retyping/ copy-pasting all this repeating code simply by replacing it with a loop:
Cbit = cell(1, 8); %always a good idea to preallocate although in your case won't make a difference
for bit = 1:8
Cbit{bit} = bitget(C, bit); figure, imshow(logical(Cbit{bit})), title(sprintf('Bit plane %d', bit));
end
Isn't that easier?
As for putting the image back together, simply use the inverse operation. You use bitget to get the bits, use bitset to set them back:
newC = zeros(size(C), class(C));
for bit = 1:8
newC = bitset(newC, bit, Cbit{bit});
end
Another option to reconstruct that image is to do the arithmetic yourself. One of the many ways (in R2016):
newC = sum(cat(3, Cbit{:}) .* permute(2.^(0:7), [1 3 2]), 3)
