Tongue of wild bee (Halictus sp)

[gpmatthews]

Having seen Cactusdave's butterfly tongue, I thought it was time for some of my own experimentation with MS ICE. The resulting image is nowhere near so spectacular, but has served to highlight a number of points.

Here is an overview image of the subject:


Microscope: Zeiss Standard
Ocular: Zeiss KPL-W 10/18
Objective: Zeiss 1.6/0.03 - 5/0.1
Substage: Leitz ICT, top lens flipped out, standard brightfield (no DIC prism)
Stack of 8 images, Zerene

I prepared a photosynth panorama using 309 images in 24 stacks with MS ICE. You can see the result at:

http://photosynth.net/view.aspx?cid=1b7 ... 40143d9c73

The software, although capable of assembling a planar panorama, will not permit upload of these and insists on only permitting upload of images stitched assuming rotating motion, which is clearly incorrect for a microscope stage. Also, it orphans blank areas of background, in which it cannot identify overlapping detail. This means I could not prepare an aesthetically acceptable rectangular image. Of course these areas can be filled in manually in a conventional image, but this doesn't assist with the Photosynth composite.

I'd be interested to know if Cactusdave has found suitable workarounds for these problems.

In all other respects MS ICE is most impressive and produces pretty good stitching where there is adequate overlap of images. You will find one or two stitching anomalies, however. Some of these are due to slight stacking banding, or other imperfections in the component stacks (i.e. my own fault!).

All 309 images were pre-processed by with a dust-map in Helicon Filter - Rik, we briefly discussed this functionality for Zerene a long time ago: any further thoughts? Filtering certainly eliminated dust bunnies.

Finally - has anyone considered using the 3D capabilities of Photosynth with a suite of stacked images from all around a specimen?

[Mitch640]

The image is pretty cool, but I barely understood the rest of it, and I did see and read the butterfly tongue thread.

[Cactusdave]

That's a very impressive and enjoyable micropanorama, and a lot of work. It is difficult to get oblique illumination to work evenly over a subject this large, but I think normal brightfield does lack something for an image like this. The absolute ideal would be a low power DIC objective, say X4. Did you look at a darkfield image, or maybe crossed or partially crossed polars?

Regarding Microsoft Photosynth, I agree it is not ideal to host micropanoramas, but it does have some distinct advantages. It is free, has a generous storage limit of 20 GB, doesn't impose advertising, has good zoom and image navigation tools and has clear copyright and image control features. I have not found a way to persuade it accept planar image acquisition as an option for upload, but have found that if cylindrical (horizontal) projection is chosen as the upload option and the field of view is set to a fairly low number, 20 degrees for example, the resultant image is very acceptable. I recommend using the Photosynth plugin for Photoshop (including Photoshop Elements) http://research.microsoft.com/en-us/dow ... d5cd04e36/ as this gives easy integration of Photosynth upload into Photoshop workflow and provides a simple work around for the orphan blank areas issue you describe.

Once you have exported your Microsoft Image Composite Editor (ICE) composite (preferably as a TIFF image) you can import it into Photoshop. Here there are options to get rid of the orphan areas. The simplest is to use the Photoshop fill or brush tools to cover the orphan areas with background colour selected using the colour picker tool in the usual way. When all edits have been done to your satisfaction then you can resize if you wish, I usually do, and export direct to Photosynth using the plugin, or using the Photosynth uploader http://photosynth.net/create.aspx bypassing Microsoft ICE. Forum user elf has shown a nice stacked micropanorama using Photosynth http://www.photomacrography.net/forum/v ... hp?t=14024 and may have some additional comments on its use.

[gpmatthews]

Thanks for these comments - very helpful. I will implement a version of your suggested workflow for my next attempt! The suggestion to use cylindrical projection sounds eminently sensible.

[gpmatthews]

- a further comment that might not have been evident; the Photosynth image in fact is taken using DIC with a 25/0.55 Leitz objective.

I was also interested in the anatomical details of the "business end" of the tongue and how the anatomy relates to the type of fluid the tongue handles: both butterfly and bee deal in nectar, but have very different tongues. The other familiar tongues are those of blow flies and house flies and these have a different capillary arrangement again.

[rjlittlefield]

Rik, we briefly discussed [dust map] for Zerene a long time ago: any further thoughts? Filtering certainly eliminated dust bunnies.

It's still "on the list", no schedule.

--Rik

[gpmatthews]

I was also interested in the anatomical details of the "business end" of the tongue and how the anatomy relates to the type of fluid the tongue handles: both butterfly and bee deal in nectar, but have very different tongues. The other familiar tongues are those of blow flies and house flies and these have a different capillary arrangement again.

further to this, see:

http://www.bbc.co.uk/news/science-environment-15068454

[Ecki]

impressive!