Figure 6
From: How levelling and scan line corrections ruin roughness measurement and how to prevent it
Numerical results for 2D polynomial background removal from image with Gaussian and exponential ACFs. Upper: Bias, plotted as measured \({\hat{\sigma }}^2\) divided by true \(\sigma ^2\) for several polynomial degrees. Lower: Standard deviation of measured \({\hat{\sigma }}^2\) (for one image), relative to the value.
