Algae, diatoms and an euglenoid

pwnell · Post by **pwnell** » Thu Aug 01, 2013 1:50 pm

20130718-DSLR_IMG_0129.jpg by pwnell, on Flickr
Diatoms on Bryopsis, 40x/0.6, FLUO-C4

20130720-DSLR_IMG_0024.jpg by pwnell, on Flickr
Crescent beach sample - red algae, 40x, FLUO-C4

20130720-DSLR_IMG_0025.jpg by pwnell, on Flickr
Crescent beach sample - red algae, 40x, FLUO-C4

20130720-DSLR_IMG_0120.jpg by pwnell, on Flickr
Crescent beach sample - red algae, , 40x, FLUO-C4

20130720-DSLR_IMG_0199.jpg by pwnell, on Flickr
GTF sample - Euglenoid, 40x, DIC

Jacek · Post by **Jacek** » Thu Aug 01, 2013 11:10 pm

You always have interesting photos, the most I like the first and last

canonian · Post by **canonian** » Thu Aug 01, 2013 11:58 pm

Jacek wrote:You always have interesting photos...

I'll second that, Waldo. Your photo's are simply amazing.
The first four could act as backgrounds in a movie like "Fantastic Voyage"

pwnell · Post by **pwnell** » Fri Aug 02, 2013 10:14 am

Thanks for the kind words. With the microscope I do feel sometimes like I am cruising on a miniature sub exploring unseen worlds. I like the surprises best - putting something under the microscope and start poking at it with light of different characteristics until something magical happens.

In the first photo I expected to only photograph Bryopsis. But once under fluorescence I could see that there are many, many different organisms living together.

The second last photo was also quite cool because a small crustacean was sitting on the right branch and it seemed like it was consuming some bits of the algae - and I think that caused the discolouration.

flyer2o12 · Post by **flyer2o12** » Sat Aug 03, 2013 12:56 pm

Very beautiful images Waldo.

You should invest in a deconvolution package, such as Autoquant, which will dramatically improve the quality of your images by removing the out of focus haze.

pwnell · Post by **pwnell** » Tue Aug 06, 2013 5:43 pm

I have used AutoQuant to 2D deconvolve the third image. It is hard to get it right as I need to find a pleasing balance between aesthetics and detail. Still working on getting better results.

discomorphella · Post by **discomorphella** » Tue Aug 06, 2013 11:38 pm

You should try 3D deconvolution, using a stack of images. Using 3D MLE you can get some incredibly sharp images, but it is pretty computationally intense, you'll want a decent gaming system effectively to run it. I've had good results using Autoquant 3D on a variety of "thick" samples, both for biological fluorescence and materials applications. You can also experiment with up to 3X superresolution as well.

David

pwnell · Post by **pwnell** » Wed Aug 07, 2013 3:24 pm

I do have a powerful system - 32GB RAM, 6 core i7-3930K CPU, SSD and nVidia GeForce 690 GPU.

My bigger issue is that I just do not see (yet) any images rendered better after deconvolution compared to what I have done using normal stacking. And for those images that do come out sharper, I am not convinced they are aesthetically more pleasing. Maybe scientifically more accurate - but not prettier.

discomorphella · Post by **discomorphella** » Wed Aug 07, 2013 4:40 pm

The concern about the aesthetics is absolutely true. You may get a much sharper image from a limited focal plane that is not as pleasing an image as the one with more background fluorescence, since we are not trying to see changes in neurons as the result of an experiment so much as take an interesting picture. What options have your tried for 3D? There are quite a few variables that go into a model-based max likelihood reconstruction, and I've had results ranging from spectacular to complete, erm, "garbage". I will go back and look over my notes.

David

RogelioMoreno · Post by **RogelioMoreno** » Fri Aug 09, 2013 1:19 pm

Waldo, nice set!

David, I would like to here more about your Autoquant´s notes.

Rogelio