Auto Fluorescence + 3D Deconvolution - Set 1

RogelioMoreno · Post by **RogelioMoreno** » Mon Sep 23, 2013 7:05 pm

I have been testing the trial versions of cellSens and Autoquant X3 to try to get better resolutions from my fluorescence stacks.

With each specimen stack you have to try all the projections that the software gives (Sum, Max, Mean, EFI) and you as artist have to choose the one that you like more. Not all the specimen gives better stack than the one you can get with stack software as Zerere or Helicon.

The following is the first set of the results. In the coming days I will publish the beauties set (Micrasterias, Closterium and Pediastrum).

Spirogyra by Rogelio Moreno G., on Flickr

Butterfly scale by Rogelio Moreno G., on Flickr

Rotifer by Rogelio Moreno G., on Flickr

Phacus by Rogelio Moreno G., on Flickr

Rogelio

Litonotus · Post by **Litonotus** » Mon Sep 23, 2013 9:44 pm

first one is outstanding! I can hardly wait for more.

pierre · Post by **pierre** » Mon Sep 23, 2013 10:42 pm

Impressive results Rogelio

Marek Mis · Post by **Marek Mis** » Mon Sep 23, 2013 11:25 pm

Very interesting and impressive images !

Marek

Charles Krebs · Post by **Charles Krebs** » Tue Sep 24, 2013 12:47 am

Yes, impressive. The spirogyra is beautiful.

arturoag75 · Post by **arturoag75** » Tue Sep 24, 2013 1:18 am

wow

René · Post by **René** » Tue Sep 24, 2013 1:26 am

Stunning indeed, Rogelio!

René

Fredlab · Post by **Fredlab** » Tue Sep 24, 2013 1:50 am

First one is incredible.
Thanks for sharing.

RogelioMoreno · Post by **RogelioMoreno** » Tue Sep 24, 2013 3:57 am

Thank you very much for your comments-

Rogelio

Jacek · Post by **Jacek** » Tue Sep 24, 2013 5:29 am

pwnell · Post by **pwnell** » Tue Sep 24, 2013 11:55 am

These are really good. How large image do you send to these applications? I found that Autoquant takes approximately 8 hours to deconvolve using spherical aberration detection a set of 40 x 110MB TIFF images at 5400 x 3500 pixels - this on a machine with 28GB RAM and fast 3.5GHz i7 processor.

Also, cellSens always crashed on me if any image stack had images larger than about 2048 x 2048.

RogelioMoreno · Post by **RogelioMoreno** » Wed Sep 25, 2013 3:26 am

Thank you for your comments.

Waldo, I have the same memory problem. Sometime cellSens crashed with Internal error. To avoid the memory problem I send 8 bits tif (instead of 16bits tif) to AutoQuant, then I crop the area of interest and finally a resize of the image (only if the amount of memory available is not enough to do the stack).

My dataset are normally around 2000 x 2000 x 24 slides.

Rogelio

Pau · Post by **Pau** » Wed Sep 25, 2013 3:35 am

Rogelio, Waldo,

Have you tried deconvolution with no fluorescence images like DIC or darkfield?. It may be great if it works

RogelioMoreno · Post by **RogelioMoreno** » Wed Sep 25, 2013 3:52 am

Pau,

I did try some DIC and polarized and I did not like the results (to ugly for my taste), I also did try a darkfield (I like more the normal stack version); but I have to try more darkfield because darkfield is like fluorescence.

Rogelio