Hi,
In ABE, PI auto-calculates the background model using polynomial surface of the order specified by the user. In DBE, the model is a spline. I think it will be very useful to also allow polynomial surface in DBE. This is useful when there are only very few background regions not contaminated by nebula emission. In such a case, ABE can easily be confused by true background and faint nebulas. This is not a problem in DBE, since the user makes the decision on what is background and what is not. However, when there are only few sample points, spline sometimes produces a model that has too many degrees of freedom. For example, if the image is properly flat-fielded, then what's left is a simple sky gradient. For such a gradient, a simple 1st-order or at most 2nd-order polynomial will be sufficient, and the determination of such a surface does not require too many sample points.
It will be nice to have such option in DBE.
Cheers,
Wei-Hao
