Is an empty field the best reference to correct vignetting in microscopy?

The vignetting effect is the radial attenuation of light intensity from the center of optical axis to edges. The most common approach to compensate for vignetting is to build the vignetting curve from an empty scene (empty field). This approach works in case the linearity of the acquisition system is a reliable approximation of reality. This work aims at identifying possible situations where the empty field could not be the best choice. We take into consideration histological samples and cell cultures and measure vignetting correction built from empty field and from the background of each specimen, that in these cases is considered quite homogeneous. We compare the different vignetting curves with the empty field one and measure the differences. Besides, we apply the vignetting correction to build a mosaic and measure which curve yields the best outcome. The experimental results prove that for a wide class of samples small non-linearities hint at using a vignetting function built directly from the specimen itself.