Scaling 8-bit images does not always retain the min,max contrast setting

[Stein Rørvik]

I am mostly working with micro-CT images, so the min/max contrast setting is very important as it defines the range of interest when applying LUTs and exporting PNGs for use in reports.

I have this issue:

The command run("Scale...", ...
does not retain the min,max contrast setting for 8-bit stacks if the Z-dimension is changed.

It works okay for 16 bit stacks, and it works okay if the Z scale is kept at 1.0
The problem occurs both in macros and if the Scale... command is called from the UI.

Here is a macro reproducing the problem:

//first scale a 16-bit image

run("T1 Head (2.4M, 16-bits)");

setMinAndMax(0, 256);

run("Scale...", "x=.5 y=.5 z=.5 width=128 height=128 depth=64 interpolation=Bilinear average process create title=t1-head-16bit-scaled.tif");

setSlice(32);

//result is ok


//then scale a 8-bit image in XY, keeping Z dimension

run("T1 Head (2.4M, 16-bits)");

setSlice(64);

run("8-bit");

setMinAndMax(0, 128);

run("Scale...", "x=.5 y=.5 z=1.0 width=128 height=128 depth=129 interpolation=Bilinear average process create title=t1-head-8bit-xy0.5-z1.0.tif");

setSlice(64);

//result is ok


//last, scale a 8-bit image in all XYZ dimensions

run("T1 Head (2.4M, 16-bits)");

setSlice(64);

run("8-bit");

setMinAndMax(0, 128);

run("Scale...", "x=.5 y=.5 z=.5 width=128 height=128 depth=64 interpolation=Bilinear average process create title=t1-head-8bit-xyz0.5.tif");

setSlice(32);

//result is wrong, the min-max settings are lost


run("Tile");



Stein

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[Wayne Rasband-2]

On Feb 16, 2017, at 6:47 AM, Stein Rørvik <[hidden email]> wrote:
>
> I am mostly working with micro-CT images, so the min/max contrast setting is very important as it defines the range of interest when applying LUTs and exporting PNGs for use in reports.
>
> I have this issue:
>
> The command run("Scale...", ...
> does not retain the min,max contrast setting for 8-bit stacks if the Z-dimension is changed.

This bug is fixed in the latest ImageJ daily build (1.51k11). Thanks for providing a macro that reproduces the problem.

-wayne

> It works okay for 16 bit stacks, and it works okay if the Z scale is kept at 1.0
> The problem occurs both in macros and if the Scale... command is called from the UI.
>
> Here is a macro reproducing the problem:
>
> //first scale a 16-bit image
>
> run("T1 Head (2.4M, 16-bits)");
>
> setMinAndMax(0, 256);
>
> run("Scale...", "x=.5 y=.5 z=.5 width=128 height=128 depth=64 interpolation=Bilinear average process create title=t1-head-16bit-scaled.tif");
>
> setSlice(32);
>
> //result is ok
>
>
> //then scale a 8-bit image in XY, keeping Z dimension
>
> run("T1 Head (2.4M, 16-bits)");
>
> setSlice(64);
>
> run("8-bit");
>
> setMinAndMax(0, 128);
>
> run("Scale...", "x=.5 y=.5 z=1.0 width=128 height=128 depth=129 interpolation=Bilinear average process create title=t1-head-8bit-xy0.5-z1.0.tif");
>
> setSlice(64);
>
> //result is ok
>
>
> //last, scale a 8-bit image in all XYZ dimensions
>
> run("T1 Head (2.4M, 16-bits)");
>
> setSlice(64);
>
> run("8-bit");
>
> setMinAndMax(0, 128);
>
> run("Scale...", "x=.5 y=.5 z=.5 width=128 height=128 depth=64 interpolation=Bilinear average process create title=t1-head-8bit-xyz0.5.tif");
>
> setSlice(32);
>
> //result is wrong, the min-max settings are lost
>
>
> run("Tile");
>
>
>
> Stein

... [show rest of quote]

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html