Re: Re: Acquiring Dental RVG on Linux
Am Fri, Jan 15, 2021 at 12:16:37PM +0530 schrieb Sonali Warunjikar:
> For mid to high exposures the quality is matching that of the vendor
> application. For low expsoure the vendor app does much better.
Would you be willing to post a vendor app image and an
Elephant version thereof ?
Maybe someone sees right away what might be the issue.
I wouldn't be surprised either if the commercial app does
some non-standard, proprietary magic to low-exposure images,
such as sharpening, denoising, extrapolating pixels,
averaging, you name it.
> Something
> coarse seems to be missing. Something like when I was assuming 16 bit
> image it was coming darker and after Aaron pointed out it was 12 bit, the
> inversion was corrected and it started coming better. Something as coarse
> as that seems missing.
That line of reasoning seems valid. I'd still be interested
in seeing the output of
$> exiftool -g1 -ee -m -u $A_VENDOR_IMAGE
> But a counter argument to that is why only for low
> exposure images.
Because, perhaps, it learns of low-exposure from the sensor,
and then applies some special treatment. The terms of
exposure might even be reflected in the DICOM metadata. And
if they are, they need to come from somewhere, which would
nearly prove that the app learneth something off the device
at runtime.
Karsten
--
GPG 40BE 5B0E C98E 1713 AFA6 5BC0 3BEA AC80 7D4F C89B
Reply to: