Use of LUTS give different results

[Robert Lockwood]

Thanks to Wayne and Jerome I have a working (up to a point) section of Java
code that creates an 16 bit gray scale imagePlus from my raw infrared
camera data and saves it as a PNG.  This image displays correctly in Fiji.

I next load a custom LUT, 256 byte each for RBB, and save that as a PNG.

If I load the gray scale PNG into Fiji and apply the custom LUT it displays
as I expect (the low values are in tans and the high in red to orange to
yellow to white).

The other PNG, with the LUT applied in code, is mostly all red.

For technical reasons the raw data start with DN values around 15,000+.
The target is a black body that I can heat to a precise temperature (I
don't go above 1,250 C in tests).  DN values at the high end are about
22,000+ with a range less than 2^14.

Why do the gray scale PNG with the LUT applied post processing and the PNG
with the LUT applied in code differ and what can I do about it? Is it a
max/min or a stretch problem?

I can attach examples if that's allowed by the Interest Group rules.  My
Java code is below:

Nate

ImagePlus imp = IJ.createImage("Untitled", imageWidth, imageHeight, 1
,imageBitsPerPix ); // 16
ShortProcessor ip = (ShortProcessor) imp.getProcessor();
short[] data = (short[]) ip.getPixels();
imageByteBuffer.position(0);
int index = 0;
while(imageByteBuffer.hasRemaining())
data[index++] = imageByteBuffer.getShort();
IJ.saveAs(imp, "PNG", imageFullPathName); // looks OK

ip.setLut(Shared.getLUT());
imp.setProcessor("LUT", ip);
IJ.saveAs(imp, "PNG", imageDisplayFolderPath);


--
When I was 12 I thought I would live forever.
So far, so good.

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

[Michael Schmid]

Hi Robert,

if you have 16-bit data and a LUT, it depends on the min & max of the display range of the pixels how it appears on the screen. By default, the display range is the full 16-bit range, 0-65535

See ImageProcessor.setMinAndMax(double min, double max)

In your case, with values between 15000 and 24000, just use

  ip.setMinAndMax(15000, 24000);


Michael
________________________________________________________________
On Dec 2, 2014, at 23:33, Robert Lockwood wrote:

> Thanks to Wayne and Jerome I have a working (up to a point) section of Java
> code that creates an 16 bit gray scale imagePlus from my raw infrared
> camera data and saves it as a PNG.  This image displays correctly in Fiji.
>
> I next load a custom LUT, 256 byte each for RBB, and save that as a PNG.
>
> If I load the gray scale PNG into Fiji and apply the custom LUT it displays
> as I expect (the low values are in tans and the high in red to orange to
> yellow to white).
>
> The other PNG, with the LUT applied in code, is mostly all red.
>
> For technical reasons the raw data start with DN values around 15,000+.
> The target is a black body that I can heat to a precise temperature (I
> don't go above 1,250 C in tests).  DN values at the high end are about
> 22,000+ with a range less than 2^14.
>
> Why do the gray scale PNG with the LUT applied post processing and the PNG
> with the LUT applied in code differ and what can I do about it? Is it a
> max/min or a stretch problem?
>
> I can attach examples if that's allowed by the Interest Group rules.  My
> Java code is below:
>
> Nate
>
> ImagePlus imp = IJ.createImage("Untitled", imageWidth, imageHeight, 1
> ,imageBitsPerPix ); // 16
> ShortProcessor ip = (ShortProcessor) imp.getProcessor();
> short[] data = (short[]) ip.getPixels();
> imageByteBuffer.position(0);
> int index = 0;
> while(imageByteBuffer.hasRemaining())
> data[index++] = imageByteBuffer.getShort();
> IJ.saveAs(imp, "PNG", imageFullPathName); // looks OK
>
> ip.setLut(Shared.getLUT());
> imp.setProcessor("LUT", ip);
> IJ.saveAs(imp, "PNG", imageDisplayFolderPath);
>
>
> --
> When I was 12 I thought I would live forever.
> So far, so good.
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html

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

[Robert Lockwood]

Thanks, Michael, that worked.  I decided to use histoMin andhistoMax for
now.

Nate

On Thu, Dec 4, 2014 at 2:21 AM, Michael Schmid <[hidden email]>
wrote:

> Hi Robert,
>
> if you have 16-bit data and a LUT, it depends on the min & max of the
> display range of the pixels how it appears on the screen. By default, the
> display range is the full 16-bit range, 0-65535
>
> See ImageProcessor.setMinAndMax(double min, double max)
>
> In your case, with values between 15000 and 24000, just use
>
>   ip.setMinAndMax(15000, 24000);
>
>
> Michael
> ________________________________________________________________
> On Dec 2, 2014, at 23:33, Robert Lockwood wrote:
>
> > Thanks to Wayne and Jerome I have a working (up to a point) section of
> Java
> > code that creates an 16 bit gray scale imagePlus from my raw infrared
> > camera data and saves it as a PNG.  This image displays correctly in
> Fiji.
> >
> > I next load a custom LUT, 256 byte each for RBB, and save that as a PNG.
> >
> > If I load the gray scale PNG into Fiji and apply the custom LUT it
> displays
> > as I expect (the low values are in tans and the high in red to orange to
> > yellow to white).
> >
> > The other PNG, with the LUT applied in code, is mostly all red.
> >
> > For technical reasons the raw data start with DN values around 15,000+.
> > The target is a black body that I can heat to a precise temperature (I
> > don't go above 1,250 C in tests).  DN values at the high end are about
> > 22,000+ with a range less than 2^14.
> >
> > Why do the gray scale PNG with the LUT applied post processing and the
> PNG
> > with the LUT applied in code differ and what can I do about it? Is it a
> > max/min or a stretch problem?
> >
> > I can attach examples if that's allowed by the Interest Group rules.  My
> > Java code is below:
> >
> > Nate
> >
> > ImagePlus imp = IJ.createImage("Untitled", imageWidth, imageHeight, 1
> > ,imageBitsPerPix ); // 16
> > ShortProcessor ip = (ShortProcessor) imp.getProcessor();
> > short[] data = (short[]) ip.getPixels();
> > imageByteBuffer.position(0);
> > int index = 0;
> > while(imageByteBuffer.hasRemaining())
> > data[index++] = imageByteBuffer.getShort();
> > IJ.saveAs(imp, "PNG", imageFullPathName); // looks OK
> >
> > ip.setLut(Shared.getLUT());
> > imp.setProcessor("LUT", ip);
> > IJ.saveAs(imp, "PNG", imageDisplayFolderPath);
> >
> >
> > --
> > When I was 12 I thought I would live forever.
> > So far, so good.
> >
> > --
> > ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

--
When I was 12 I thought I would live forever.
So far, so good.

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