STF Auto Stretch for inverted images

[Yuriy Toropin]

Hi,

Looks like default settings of STF Auto Stretch (-1.25, 0.25) while working good for "normal" images don't do the job right for the "inverted" ones. "Auto Stretched" inverted image looks very contrast with background level set too high (see screenshot attached, the contrast one in below is the "auto stretched" invert image, while the one higher has been stretched manually).

STF Auto Stretch is a very handy feature, could it be better adopted for inverted images?

The purpose - fainter details are better visible on inverted images, helpful for inspection during processing.

Thanks a lot!

PS: Juan, sorry if I didn't get any undocumented essencials of PI again... 

[georg.viehoever]

Yuriy,

maybe doing a Process/IntensityTransformations/Invert before you do the STF resolves the issue for you?

Georg

[Yuriy Toropin]

Hi Georg,

I'm inverting grayscale image by simply selecting Image->Invert (or by shotcut Ctrl+I).
As far as I can see from Processing Console, it's the same process as that one under Process/IntensityTransformations/Invert.

STF Auto Stretch applied after it with very contrast result (with default parameters in the dialog form).

Sorry, George, may be I don't fully understand what did you mean...

Juan, just idea - may be add a checkbox "Negative" in STF Auto Stretch dialog so it will be possible to see negative stretched screen representation without really inverting teh image for this?

[georg.viehoever]

Hi Yuriy,

I probably misunderstood you. If you just want to look at an image with dark background in an inverted way, adding a "Show Negative" checkbox to STF might be a good idea. Sometimes this reveals details otherwise hidden.

However, if you prefer to work with images that actually have a light background (high numeric values) and dark stars (low numeric values), you will probably find that a number of PI tools are working with the assumption that the background is usually dark (e.g. histogram transform, curves transform, probably de-ringing, ACDNR, ...). Adding an "inverted" checkbox to all those tools might not be such a good idea.

Georg

[Juan Conejero]

Hi Yuriy and Georg,

Of course count with this feature implemented in PI 1.5.3. Thank you for your help and ideas!