Adjust contrast

Dear all,
I am processing some of my images. Normally I start aplying some gaussian filtering, then I subtract the spatial variations of the background intensities using rolling ball function and afterwards I subtract some background signal by thresholding it.
In some of my images that have very low signal, I have observed, that if before all of that, I enhance contrast “auto” and apply LUT, I got much better imaging results (the signal that comes from my cells is better distinguised from the background noise) than when I do not do it. But I cannot understand why. Does anyone have any idea? Is it correct to apply auto contrast enhancement and apply LUT at the beginning of the processing?

Thanks,
Miguel

If area selection is your goal, there is no problem. If you want to do quantitative intensity measurements, you have just ruined your data.

When an area selection (segmentation) is necessary to measure intensity in a certain area only, find that area using a duplicate (of the original image) and use whatever means are necessary, then collect the intensities from the original masked by the duplicate.

Thanks for the advice, I appreciate. This is just for visualization purpose, not quantification, so I think it is find if I do it, right?