GradientCorrection challenge

Hi Juan,

thanks for the clarification and additional information. This sounds like a great addition!

CS Gerrit
 
Juan Conejero said:
As Vicent says, this new feature is similar to ABE in the device we use to perform a two-dimensional interpolation of the simplified gradient function. However, this interpolation is guided by a multiscale analysis instead of the simplistic statistical model implemented in the ABE tool. This allows us to perform an initial robust simplification of the gradients in the image, which greatly facilitates the work of the core gradient modeling and correction algorithm, making it more efficient and controllable. The new simplification feature applies to complex gradient cases where no bright objects cover a significant area of the image.
Click to expand...
Previously, having seen only the screen shot earlier in this thread, I thought clicking "simplified model" would skip the usual model generation, and if a user wanted to first do a simplified model, then the original gradient correction, they would have to run the process twice.

Now that I see the example videos, I am getting the impression that the "simplified model" is a preprocessing step to the previously-existing model generation, which will still execute. If that interpretation is correct, it could be useful to signal that more clearly. Perhaps the panel could be named "simplified-model preprocessing" or something like that, and be placed above the "model generation" panel, since it happens first.
 
Once I started doing a lot of post-processing on starless images, I also switched to starless when applying DBE. Is there any advantage, or disadvantage, to using GC on starless images?

Cheers,
Scott
 
Scott Badger said:
Once I started doing a lot of post-processing on starless images, I also switched to starless when applying DBE. Is there any advantage, or disadvantage, to using GC on starless images?

Cheers,
Scott
Click to expand...
No advantage at all, and many potential disadvantages. Stars are part of the image data that GradientCorrection uses to compute gradient models internally. While stars are small-scale structures "invisible" to large-scale analysis, their influence on adjacent regions, such as star halos or PSF tails, can be important to the accuracy of computed models. Star removal tools tend to generate small-scale gradients and other artifacts since they are actually automated clone stamps or inpainting tools. Working with complete images is the recommended practice to achieve reliable gradient corrections with the GradientCorrection process.
 
johnpane said:
Previously, having seen only the screen shot earlier in this thread, I thought clicking "simplified model" would skip the usual model generation, and if a user wanted to first do a simplified model, then the original gradient correction, they would have to run the process twice.

Now that I see the example videos, I am getting the impression that the "simplified model" is a preprocessing step to the previously-existing model generation, which will still execute. If that interpretation is correct, it could be useful to signal that more clearly. Perhaps the panel could be named "simplified-model preprocessing" or something like that, and be placed above the "model generation" panel, since it happens first.
Click to expand...

Simplified models are computed and applied during the preprocessing phase before the main gradient modeling and correction algorithm. Both tasks are not independent or simply sequential, but tightly tied together, since simplification changes the way the gradient correction algorithm works in a significant way.

In our opinion, the current tool layout is the most appropriate: first, the main algorithm parameters, then the auxiliary, optional task parameters.
 
You must log in or register to reply here.
Back
Top