Skip to main content
Participant
September 23, 2018
Answered

Remove uneven illumination (e.g. in microscope images)

  • September 23, 2018
  • 1 reply
  • 1245 views

Dear wise people of the internet!

I took photos with a brightfield microscope for scientific use, but some lenses have very specific "vignettes" because they haven't been properly cleaned for years. The first image you can see is what my specimen looks like:

You can see that the top image is disturbed by lens-specific "noise". I took a blank image of a white background as can be seen below:

Now, I know there's a rather simple way or getting rid of this uneven illumination and to make a perfectly nice image that looks more like this (no vignette, no yellow coloring, no color gradient etc):

Only problem is that I forgot how to do it ...but it definitely utilizes the blank image to remove this vignette. I tried "Apply Image" or "Subtract" or "Difference" but I only end up getting black spots where there's supposed to be a white background. In the method I'm trying to remember, there was no use of any blur or Photoshop filters, it was really just somehow based on removing this blank.

I would really appreciate some help with this Thank you!

This topic has been closed for replies.
Correct answer Chuck Uebele

Off-hand, I would say that if you have a blank image, take that along with your real image and put them into Photoshop. Invert the blank image and play with the blend modes to cancel out the brightness and color.

Or

1 reply

Chuck Uebele
Community Expert
Chuck UebeleCommunity ExpertCorrect answer
Community Expert
September 23, 2018

Off-hand, I would say that if you have a blank image, take that along with your real image and put them into Photoshop. Invert the blank image and play with the blend modes to cancel out the brightness and color.

Or

Participant
September 23, 2018

That is the first thing I tried, but no, the method had a few more steps to get the final image.

Chuck Uebele
Community Expert
Community Expert
September 23, 2018

Not knowing that the specimen looked like, the blend with Linear Dodge seemed to cancel out the variations fairly well. Of course you could play with the opacity of that layer too. Any idea what additional steps were in the method you're thinking about?