Estimates non-uniform background level by fitting a parametric surface to background samples.

This ImageJ plugin allows removal of non-uniform background noise. It finds a least-squares fit of background samples within the image to one of two intensity profiles: 1) a plane, or 2) a 2D cubic polynomial. The estimated background is subtracted from the input image. This plugin can operate on either the current image in a stack or on an entire stack. In the case that it operates on the stack, background fitting operates on a slice-by-slice basis. Additionally, the background estimate image may be created to inspect the quality of the fit.

Download this plugin here.