imadjust low_in high_in, why only in the range [0,1] ?

Hello, it looks a bit strange that the function imadjust accepts, other than single/double numeric types, also integer types, but actually one is forced to put low_in and high_in values only in the range [0,1]. It seems one forcedly have to convert its input image to double (so increasing a lot memory usage) to adjust an integer valued image.

Am I wrong?

Max

Answer 1

[Walter Roberson]

You have misunderstood. You do not need to change the input image datatype. What you are looking at is the contrast limit specification, and contrast is always a relative factor, from no contrast permitted (0) to comple contrast permitted (1)

[Massimo Zanetti]

Thanks Walter, but I still do not understand very well...

Assume A is a matrix of uint16 values. So in the range [0, 65535]. I want to adjust the image by mapping values between [600, 30000] to [0, 65535] and clip <600 and >30000. How am I supposed to do?

[Massimo Zanetti]

By reading documentation:

J = imadjust(I,[low_in high_in],[low_out high_out]) maps intensity values in I to new values in J such that values between low_in and high_in map to values between low_out and high_out.

It seems it works as I would imagine, but of course this cannot happen because

J = imadjust( A , [600, 30000] , [0, 65535] )

returns error!

[Walter Roberson]

J = imadjust( A , [600, 30000]/intmax(class(A)) , [0, 65535]/intmax(class(A)) );

Though you would probably use

J = imadjust(A, [600, 30000]/65535, [0 1]);

[Massimo Zanetti]

Thank you Walter.

I came up with the same. However, I think the documentation is misleading on how to use this function, simply because it is not true that values in I are mapped, if you have to scale them before mapping :)

Thanks again.