I know that similar suggestions have been made in the past, but bear with me.
The basic idea would be to let the user select, with a lasso-type tool, several/most/all of the areas of an image containing only noise (and thus usable for per-image-based noise profiling).
From previous answers to this, it seems that Neat Image noise profiling algorithms cannot use non-rectangular areas, and therefore this suggested feature would be pointless.
But what about making Neat Image analyze this/these non-rectangular selected area(s) to find the largest, continuous, rectangular-shaped areas within them, and then use those rectangular areas found for the actual profiling?
Automatic, non-rectangular area selection noise profiling
Neat Image does that automatically when you use Auto Profile. It asks you to verify that the main area is good enough for profiling. If it is good then all other areas found by NI automatically will be good as well. It doesn't swow them all (there are many) to save your time.
Thank you for the suggestion!
Kind regards,
Vlad
Thank you for the suggestion!
Kind regards,
Vlad
Thank you for your answer, Vlad.
However, if Automatic Profiling was always the best possible way to profile the image, then there would be no actual need for the current implementation of manual profiling, either.
This suggestion was meant as an alternative way for these manual profiling scenarios, when Auto Profiling doesn't work for the user and for the image being processed, and he wants to perform manual profiling but without quite the hassle of trying to manually find the largest rectangle of analysis rectangles that will fit in each featureless noisy area.
Anyway, and again, thanks again for your answer and interest in the suggestion. It was just an idea, lol![Laughing :lol:](./images/smilies/icon_lol.gif)
However, if Automatic Profiling was always the best possible way to profile the image, then there would be no actual need for the current implementation of manual profiling, either.
This suggestion was meant as an alternative way for these manual profiling scenarios, when Auto Profiling doesn't work for the user and for the image being processed, and he wants to perform manual profiling but without quite the hassle of trying to manually find the largest rectangle of analysis rectangles that will fit in each featureless noisy area.
Anyway, and again, thanks again for your answer and interest in the suggestion. It was just an idea, lol
![Laughing :lol:](./images/smilies/icon_lol.gif)