I use Neat Image pro to process astronomical images and find it a very interesting tools.
A few suggestions/questions :
It would be great to get support for FITS image format (FP and 16/32 bits great scale format) in input and output.
Is there a way to setup a noise profile which would be typical of scientific digital cameras where you know
readout noise
Gain (e- per ADU)
and assume a real gaussian noise distribution.
---------
Other issue is that if I open a 16 bits TIFF image, in most cases the noise level is around 10 to 30, so it is very hard to see the fainter part of the images when selecting an area for the noise estimate.
Would be great to have a contrast/brightness control to be able to adjust the image when working in Neat Image.
Other than that, I must congratulate you on your algorithm... I worked a bit on denoising algorithms and found your implementation to be spectacular !
-- Benoit Schillings
Astronomical images
Thank you for your comments and suggestions!
It seems, however, that implementing all these suggestions will amount to producing a new, specialized product to process astronomical images. - Users of digital cameras and scanners will hardly need FITS format. These devices usually do not have enough precision and control to accurately record the conditions and results of experiments/measurements. This make the domains of commercial and astronomic (scientific) imaging so different, which requires different tools. Therefore, to fully answer the needs of astronomic community, we would have to develop a separate product, which takes into account the specificity of the domain. That may be a good direction for our further work.
Regarding the contrast/brightness control, I understand that you sometimes have difficulties with discerning feature-less areas if they are dark. Currently, the only way out is to use the Negative button in NI, but of course there may be other ways, like (temporary) brightness/contrast adjustment. We will consider this, thank you for the suggestion.
Once again, thank you for your kind comments!
Vlad
It seems, however, that implementing all these suggestions will amount to producing a new, specialized product to process astronomical images. - Users of digital cameras and scanners will hardly need FITS format. These devices usually do not have enough precision and control to accurately record the conditions and results of experiments/measurements. This make the domains of commercial and astronomic (scientific) imaging so different, which requires different tools. Therefore, to fully answer the needs of astronomic community, we would have to develop a separate product, which takes into account the specificity of the domain. That may be a good direction for our further work.
Regarding the contrast/brightness control, I understand that you sometimes have difficulties with discerning feature-less areas if they are dark. Currently, the only way out is to use the Negative button in NI, but of course there may be other ways, like (temporary) brightness/contrast adjustment. We will consider this, thank you for the suggestion.
Once again, thank you for your kind comments!
Vlad
If you're looking for a specialized product to do astronomical tasks, check out ImagesPlus. It's not really a competitive product, as it specializes only in astrophotography.
www.MLUnsold.com
www.MLUnsold.com
A Wild Idea
I just thought of an idea to overcome problems to remove noise in very dark images.
The solution is simple. Invert the colours of the image before loading it in Neat Image. It works great!
It does produce somewhat different results than the regular way. At least from what I could see.. Maybe it is just me? =)
The solution is simple. Invert the colours of the image before loading it in Neat Image. It works great!
It does produce somewhat different results than the regular way. At least from what I could see.. Maybe it is just me? =)
Yes, it is similar. But is not the difference that the Negative button only visually helps the user to find suitable areas for noise reduction, while negating the image before loading it in NI changes the way noise reduction works?
For example very dark areas would be bright and vice versa. The noise may appear different and so will the noise reduction?
For example very dark areas would be bright and vice versa. The noise may appear different and so will the noise reduction?
I am not even sure that there is a better noise reduction with the negative image. I just noticed a difference in the results when I tried.
My thesis was that it could be "easier" for NI to reduce noise in brighter areas than in darker. Therefore inverting the image would allow NI to reduce more noise in the dark areas than it normally would. This would be beneficial in night and/or astronomy pictures.
However I do not know if this actually works or not. It could be worth investigating more.
My thesis was that it could be "easier" for NI to reduce noise in brighter areas than in darker. Therefore inverting the image would allow NI to reduce more noise in the dark areas than it normally would. This would be beneficial in night and/or astronomy pictures.
However I do not know if this actually works or not. It could be worth investigating more.
There is a possibility of different noise reduction in bright and dark areas related to the gamma curve, but I think a non-evenly equalized noise profile may rather be the reason for that difference in noise reduction. If the 'bright' sliders are significantly different from 'dark' ones then noise reduction applied to bright and dark image areas is naturally very different.
Vlad
Vlad