intlut() alternative for LUT consisting of doubles

Is there a fast alterative to intlut() accepting an array of doubles as LUT?

Problem is, I want to convert grayscale values (uint8) of an image using a LUT consisting of 256 doubles.

Example:

lut = (1:256)+0.1; % upshift by 0.1
A = uint8(randi(256,50));

% current approach
lutMat = repmat(reshape(lut, 1, 1, numel(lut)), ...
        size(A));
idx = sub2ind( ...
  size(lutMat), ...
  repmat((1:size(A, 1))', 1, size(A, 2)), ...
  repmat((1:size(A, 2)) , size(A, 1), 1), ...
  double(A));
B = lutMat(idx);

The above code works fine, but takes a lot of time and needs too much memory

[Jan]

When I run your code, I get an error:

 ??? Error using ==> sub2ind at 58
 Out of range subscript.

[Stefan]

I forgot to add the line

 lut = repmat(reshape(lut, 1, 1, numel(lut)), ...
          size(A));

just after

% current approach

I changed the code in the original question, accordingly.

[Image Analyst]

You're still having the problem? Didn't my code do what you want?

Answer 1

[Jan]

I'm not sure if I understand the question. Does this help:

B = lut(A);

Answer 2

[Image Analyst]

If you're just adding a constant value to the lut-transformed values, like 0.1, you can just do this:

lut = uint8(1:256); % upshift by 0.1
A = uint8(randi(256,50));
B = double(intlut(A, lut)) + 0.1;

If it's not some constant added value, then you can do it like this:

lut = 1000 * rand(256, 1); % A Random reassignment between 0 and 1000.
A = uint8(randi(256,50));

B2 = zeros(size(A), class(lut));
for inputValue = uint8(0) : uint8(255)
   indexes = (A == inputValue);
   B2(indexes) = lut(inputValue + 1);
end