Multi - Level Focused & Stacked Images
Moderators: MacroMike, nzmacro, Ken Ramos, twebster, S. Alden
Multi - Level Focused & Stacked Images
I recall seeing an advertisement somewhere on the web about multilevel focusing. Think it was Nikon. Then the other day I read that Motic's new upgrade to the interface program I am currently using provides for multiple level focus to increase clarity/dof in images. Then today, I see Charles Krebs fantasic image of "Waterflea Gladiator" made from five separate images. Stunning! So how do you guys do it ?? Is it software that does it for you ? Do you assemble the separate images manually in Photo Impact or PhotoShop? I am excited about this because I just starting learning about the use of layers in PhotoShop. It can't be as simple as putting each focused level (or optical "section") on a separate layer and then stacking them? I am sure some of you have tried it. Of course each image will have parts grossly out of focus but that can be cleared away and nothing but the sharp pieces will be left to assemble the final image.
George
George
- gpmatthews
- Posts: 539
- Joined: Tue Apr 27, 2004 11:54 am
- Location: Horsham, UK
- Contact:
There are two techniques - the long winded and the quick (but not always best).
The long winded way involves using a set of images that are combined manually by any convenient method. A little while back I posted an image of Actinoptychus prepared by this method, although I also used Method 2 to assist (http://www.photomacrography1.net/forum/ ... ghlight=).
The quick method uses software to combine a stack of images. There are some astronomical programs that can do this, but the one most likely to be used by amateur microscopists is probably CombineZ. This can be downloaded for free from Alan Hadley's website (he is the author) via a link at http://www.microscopy-uk.net/full_software.html. This picture was compiled using this software, followed by some retouching: http://www.photomacrography1.net/forum/ ... highlight=
This software can be very successful, but can also produce unsightly artefacts. Practice, as always, makes perfect.
I'm sure Charlie will have some further tricks to add...
The long winded way involves using a set of images that are combined manually by any convenient method. A little while back I posted an image of Actinoptychus prepared by this method, although I also used Method 2 to assist (http://www.photomacrography1.net/forum/ ... ghlight=).
The quick method uses software to combine a stack of images. There are some astronomical programs that can do this, but the one most likely to be used by amateur microscopists is probably CombineZ. This can be downloaded for free from Alan Hadley's website (he is the author) via a link at http://www.microscopy-uk.net/full_software.html. This picture was compiled using this software, followed by some retouching: http://www.photomacrography1.net/forum/ ... highlight=
This software can be very successful, but can also produce unsightly artefacts. Practice, as always, makes perfect.
I'm sure Charlie will have some further tricks to add...
Graham
George,
Here's a fairly lengthy reply, because I think this is an extremely useful technique.
For us folks using a "basic" compound light microscope there are essentially two methods as I see it... "automatic" using software designed for the task; and "manual" using a program like Photoshop (or other good image editor that lets you work with layers). You should be aware that even if you have the best "auto" software, you will find that you often (usually?)need to do some manual "clean-up" after running it.
The idea is straightforward in concept... you take a series of identically exposed images while you gradually move the focus from the top through to the bottom of the subject. Ideally, there should be a bit of overlap in the DOF of each shot. And ideally the subject should not move at all during the series of shots. (Which is, of course, a problem with live subjects). Then, using the computer, you take only the sharpest section of each image and combine them to form one picture with extended DOF. Sounds like it might not be too complicated, but it takes some experience and often a great deal of time to do manually, as each image may have to be "sized" slightly and/or rotated or moved a bit to line up (especially the case with stereo microscopes)
I have recently tried to find demos of as many of these software programs as I could, and do a comparison. They range in price from $0 to over $4000 . (To be fair, the hugely expensive ones are capable of all sorts of measuring, counting, and control operations that are important to high end research operations... but are of no use to me). Microscope and microscope camera manufacturers also have software to do this, but it is part of an integrated multifunction software package and I have not yet been able to try "demos" of them. (George...I think there is such a function in your Motic software... I believe they call it "amalgamation". I tried it with modest results)
I have looked at the following so far (if you hear of others, please let me know):
CombineZ4.exe...$0 (donation requested if you like it), by Alan_Hadley. alan_hadley@blueyonder.co.uk
http://www.hadleyweb.pwp.blueyonder.co. ... unning.htm
This is a free program. It is quite sophisticated and I have played with it off and on for the past year. My results have been mixed. Sometimes it looks great. Other times it required so much "clean-up" it was easier to do it all manually in Photoshop. This could be user incompetence on my part. Very slow with large file sizes. It's worth trying. The price is right.
Image in Focus...$700. Special K Software. OK, but blocky looking output.
IF Junior... about $470 www.alicona.com. This look as if might be very good. Unfortunately their demo program outputs only a very small size picture (heavily pixelated) with a huge watermark. Unfortunately it's very tough to tell how good it is from the demo, since fine detail is non-existent.
AutoMontage Essentials... about $1000. www.syncroscopy.com. An excellent program. Too bad it is so expensive. Very effective and unique editing capability for "clean-up" after running stack. The demo is good as well, but gives a very heavily watermarked image. It is still possible to see what it does. (AutoMontage Pro is much more expensive, but not really much more capable in the "combining" function itself)
Helicon Focus and Helicon Focus Pro... $115 and $250. An excellent program. www.helicon.com.ua. This is a small software company in the Ukraine. Demo is not crippled or watermarked; but time limited. This program was a surprise. Very intelligently designed, very fast operation, and perhaps the best output of the programs I tried. There is a unique "dust map" filter that can be included in the stack to eliminate recurring marks from dust on the camera sensor or microscope optic. The "Waterflea Gladiator" image I posted was put together with the demo of this program (and some Photoshop touch-up). The "Pro" version provides a sort of masking capability that I have just begun to explore. It looks like it could eliminate some "touch-up" work, but the "regular" version would be fine. (When you buy the Pro version you also get a user key for another program they produce ... Helicon Filter. I haven't looked at it much yet, but it seems to offer some unique approaches to many of the types image correction I normally do in Photoshop). As they say, I have no connection with this company at all, but was simply very excited to see such a capable program at a reasonable cost. I don't have a great deal of experience with it yet, but my early results are amazingly good.
If you want to try this "stacking" technique, you really should try the Helicon Focus demo. In fact, I would love to hear anyones experience using AutoMontage Demo and Helicon Focus on the same image stack (...Graham ... I know you have a bunch of image stacks on your hard drive). They are the two best programs I've tried so far.
Here's a fairly lengthy reply, because I think this is an extremely useful technique.
For us folks using a "basic" compound light microscope there are essentially two methods as I see it... "automatic" using software designed for the task; and "manual" using a program like Photoshop (or other good image editor that lets you work with layers). You should be aware that even if you have the best "auto" software, you will find that you often (usually?)need to do some manual "clean-up" after running it.
The idea is straightforward in concept... you take a series of identically exposed images while you gradually move the focus from the top through to the bottom of the subject. Ideally, there should be a bit of overlap in the DOF of each shot. And ideally the subject should not move at all during the series of shots. (Which is, of course, a problem with live subjects). Then, using the computer, you take only the sharpest section of each image and combine them to form one picture with extended DOF. Sounds like it might not be too complicated, but it takes some experience and often a great deal of time to do manually, as each image may have to be "sized" slightly and/or rotated or moved a bit to line up (especially the case with stereo microscopes)
I have recently tried to find demos of as many of these software programs as I could, and do a comparison. They range in price from $0 to over $4000 . (To be fair, the hugely expensive ones are capable of all sorts of measuring, counting, and control operations that are important to high end research operations... but are of no use to me). Microscope and microscope camera manufacturers also have software to do this, but it is part of an integrated multifunction software package and I have not yet been able to try "demos" of them. (George...I think there is such a function in your Motic software... I believe they call it "amalgamation". I tried it with modest results)
I have looked at the following so far (if you hear of others, please let me know):
CombineZ4.exe...$0 (donation requested if you like it), by Alan_Hadley. alan_hadley@blueyonder.co.uk
http://www.hadleyweb.pwp.blueyonder.co. ... unning.htm
This is a free program. It is quite sophisticated and I have played with it off and on for the past year. My results have been mixed. Sometimes it looks great. Other times it required so much "clean-up" it was easier to do it all manually in Photoshop. This could be user incompetence on my part. Very slow with large file sizes. It's worth trying. The price is right.
Image in Focus...$700. Special K Software. OK, but blocky looking output.
IF Junior... about $470 www.alicona.com. This look as if might be very good. Unfortunately their demo program outputs only a very small size picture (heavily pixelated) with a huge watermark. Unfortunately it's very tough to tell how good it is from the demo, since fine detail is non-existent.
AutoMontage Essentials... about $1000. www.syncroscopy.com. An excellent program. Too bad it is so expensive. Very effective and unique editing capability for "clean-up" after running stack. The demo is good as well, but gives a very heavily watermarked image. It is still possible to see what it does. (AutoMontage Pro is much more expensive, but not really much more capable in the "combining" function itself)
Helicon Focus and Helicon Focus Pro... $115 and $250. An excellent program. www.helicon.com.ua. This is a small software company in the Ukraine. Demo is not crippled or watermarked; but time limited. This program was a surprise. Very intelligently designed, very fast operation, and perhaps the best output of the programs I tried. There is a unique "dust map" filter that can be included in the stack to eliminate recurring marks from dust on the camera sensor or microscope optic. The "Waterflea Gladiator" image I posted was put together with the demo of this program (and some Photoshop touch-up). The "Pro" version provides a sort of masking capability that I have just begun to explore. It looks like it could eliminate some "touch-up" work, but the "regular" version would be fine. (When you buy the Pro version you also get a user key for another program they produce ... Helicon Filter. I haven't looked at it much yet, but it seems to offer some unique approaches to many of the types image correction I normally do in Photoshop). As they say, I have no connection with this company at all, but was simply very excited to see such a capable program at a reasonable cost. I don't have a great deal of experience with it yet, but my early results are amazingly good.
If you want to try this "stacking" technique, you really should try the Helicon Focus demo. In fact, I would love to hear anyones experience using AutoMontage Demo and Helicon Focus on the same image stack (...Graham ... I know you have a bunch of image stacks on your hard drive). They are the two best programs I've tried so far.
Last edited by Guest on Thu Mar 24, 2005 11:57 am, edited 1 time in total.
-
- Posts: 1200
- Joined: Tue Jul 13, 2004 10:50 am
- Location: Issaquah, WA USA
Hi George, hi Charlie,
for stacking, i use CombineZ and Photoshop (manually stacking) with different
results in different picture material. In Brightfield and incident light results with CombineZ are good,
Rheinberg and DIC sometimes causes artifacts. In Photoshop i have more flexibility.
Thanks Charlie for the additional informations about the different software solutions,
i will test them !
The following exaples are stackes of 16 pics each, the first made with CombineZ,
the second handmade with Photoshop,
Radiolaris shell in Rheinberg illumination, Obj. 20x, Periplan 10x, CP990
Thanks,
Gerd
for stacking, i use CombineZ and Photoshop (manually stacking) with different
results in different picture material. In Brightfield and incident light results with CombineZ are good,
Rheinberg and DIC sometimes causes artifacts. In Photoshop i have more flexibility.
Thanks Charlie for the additional informations about the different software solutions,
i will test them !
The following exaples are stackes of 16 pics each, the first made with CombineZ,
the second handmade with Photoshop,
Radiolaris shell in Rheinberg illumination, Obj. 20x, Periplan 10x, CP990
Thanks,
Gerd
Gerd... you did a beautiful job with this subject! Doing a 16 stack of a subject like this in Photoshop can make you dizzy. I love radiolarians. It's one of the reasons I went searching for stacking software. One of our members, Piotr, has some really nice radiolarian shots on his site here: http://www.pirx.com/droplet/gallery.html
You beat me by one image. My most ambitious "manual" stack was 15 images on this this cool looking Lesquereusia amoeba test:
http://www.photomacrography1.net/forum/ ... .php?t=773
I am curious.... What pixel dimensions did you use for the stack you ran through CombineZ4? The Canon 10D files are a bit over 3000x2000 pixels (18 megs). If I put 16 of them in CombineZ4 it would take an extremely long time to process.
I would love to hear your impressions of the results using the programs I mentioned. You could run this same stack of the radiolarian.
You beat me by one image. My most ambitious "manual" stack was 15 images on this this cool looking Lesquereusia amoeba test:
http://www.photomacrography1.net/forum/ ... .php?t=773
I am curious.... What pixel dimensions did you use for the stack you ran through CombineZ4? The Canon 10D files are a bit over 3000x2000 pixels (18 megs). If I put 16 of them in CombineZ4 it would take an extremely long time to process.
I would love to hear your impressions of the results using the programs I mentioned. You could run this same stack of the radiolarian.
-
- Posts: 1200
- Joined: Tue Jul 13, 2004 10:50 am
- Location: Issaquah, WA USA
Grahm, Gerd, and Charlie, Thanks all for the very informative response. Wow. So much to take in. Those are some fantastic radiolarian images Gerd. I am especially encouraged by the fact that I think the one you put together manually is better than the Combine Z version. But those artifacts around the periphery I am sure will clean up nicely in PhotoShop. Charlie, I will definately check on the helicon software. Looking at your water flea has sold the program for me! That's about the best endorsement any imaging software could ever have. And yes, I do have an "amalgation" option in my Motic software. Never tried it because I didn't know, until now, what it was for. I'll have to check it out. I been spending lots of time with PhotoShop lately. I find I enjoy processing the image almost as much as taking the image. I realize now that a good "happy snap" is just the beginning. I never had a clue what a good microscopy image was until I came here and saw what the "amateurs" were doing. Thanks for all the great information and friendly encouragement.
George
George
Stacking with IP4.5
I have been using Image Pro 4.5 (now there is a V5.1... http://www.mediacy.com/ipp/ippfeatures.htm ) which has the stacking facility they call EDF = extended depth of field. Quite good results, but as noted above sometimes spoiled by angled polygons in the background.
When I sent them a series of pics and the resulting artefact-riddled composite, Media Cybernetics told me this was a problem of variation in my lighting intensity between shots. They were partly right and proved it by making and sending me a video clip of my background intensity, which was certainly wobbling around, and I can see why the prog had trouble equalising it. I switched to a DC regulated light source and the problem nearly went away, and I am now going even further to use a feedback-based light equaliser on the scope input (not implemented yet). It seems to be very important, this constant lighting.
The IP function gives the option of stacking based on x-y contrast or z contrast, with a number of other options. It will also generate a topographic map showing in grey-scale where each bit of the final image came from... and this map makes a nice image in itself!
But... I should have thought that Photoshop, in the sense of being the best standard graphics engine, must be able to do stacking by means of a plug-in. I have tried to find one, but with no success, yet it should be a simple enough routine for someone to write, if you're good at writing routines, which I am not, the only routine that works for me is the morning coffee.
Anybody feel competent?
Someone in Australia wrote a routine for Scion Image (US-government public domain prog) and published a paper on how he did this. However he didn't include the routines themselves. And I can't find the paper now :-(
When I sent them a series of pics and the resulting artefact-riddled composite, Media Cybernetics told me this was a problem of variation in my lighting intensity between shots. They were partly right and proved it by making and sending me a video clip of my background intensity, which was certainly wobbling around, and I can see why the prog had trouble equalising it. I switched to a DC regulated light source and the problem nearly went away, and I am now going even further to use a feedback-based light equaliser on the scope input (not implemented yet). It seems to be very important, this constant lighting.
The IP function gives the option of stacking based on x-y contrast or z contrast, with a number of other options. It will also generate a topographic map showing in grey-scale where each bit of the final image came from... and this map makes a nice image in itself!
But... I should have thought that Photoshop, in the sense of being the best standard graphics engine, must be able to do stacking by means of a plug-in. I have tried to find one, but with no success, yet it should be a simple enough routine for someone to write, if you're good at writing routines, which I am not, the only routine that works for me is the morning coffee.
Anybody feel competent?
Someone in Australia wrote a routine for Scion Image (US-government public domain prog) and published a paper on how he did this. However he didn't include the routines themselves. And I can't find the paper now :-(
And another thing...
Images to be automatically stacked should not be jpegs, as these have artefacts in them which involve steps in pixel level. The program may well see these steps as "maximum local pixel-to-pixel contrast" and include the area as being in-focus, which of course it may well not be.
Solution is only use TIFF etc images, can jpeg the result afterwards.
Solution is only use TIFF etc images, can jpeg the result afterwards.
- twebster
- Site Admin
- Posts: 1518
- Joined: Mon Apr 19, 2004 10:55 pm
- Location: Phoenix "Valley of the Sun", Arizona, USA
Hi ya' guys
Just in time for this discussion, here is a news release that was attached to an email I received from Danylo Kozub, Chief Software Developer, of Helicon Co.
Best regards to all as always,
Just in time for this discussion, here is a news release that was attached to an email I received from Danylo Kozub, Chief Software Developer, of Helicon Co.
Charles sings high praise for this program in an earlier response to this thread and this release does look very interesting to me. My efforts with photoshop have been less than successful. I may just give it a try in the next few weeks.We are pleased to announce the release of Helicon Focus Pro 3.10, the
program for combining partially focused images.
Helicon Focus solves shallow DOF (depth-of-field) problem which almost
always appears in macroshooting (<10cm) and in digital
microphotography (optical microscopes). The program finds the focused
areas in each image and combines these areas into one perfectly
focused image.
Starting from v.3.00, Helicon Focus also automatically aligns images
as objects often change their size and position from shot to shot.
This function is especially important for macrophotography.
New version of Helicon Focus has enhanced algorithm and improved
performance (up to 3x) compared to 2.03 version. It also includes
"dust map" feature to remove dust and hot pixels from the resulting
image.
Pro version includes the possibility to fine tune opacity maps in
external graphic editor to fix minor artifacts. Only a few minutes of
work will result in perfect image.
Helicon Focus Pro version costs $250 and includes free license for
Helicon Filter Pro, advanced image editing software. Helicon Focus
costs $115.
Helicon Focus supports most popular file formats including all RAWs.
It has no limitation on number of images or resolution.
Helicon Focus is THE solution for photo enthusiasts looking for
infinite depth of field as well as for professionals in
microphotography.
The full featured 30 day evaluation version and the sample images are
available at the product home page:
http://heliconfocus.com/pages/?focus_overview
Direct download link:
http://downloads.heliconfilter.com/HeliconFocus3.10.exe
Best regards to all as always,
Tom Webster
Administrator
Phoenix "The Valley of the Sun", Arizona, USA
Think about this...maybe Murphy is an optimist!!!
Administrator
Phoenix "The Valley of the Sun", Arizona, USA
Think about this...maybe Murphy is an optimist!!!
Hi Charlie, hi Planapo, hi Tom
i know Imagepro too, it works well, but price for me is much to high.
I downloaded the trial version of imagepro express 5 today and computed the radiolaria stack
again. And i tested heliconfocus too, the same stack as the pix before.
You can see the results beneath.
Charlie, i usually take 45 minutes on the average to handle about 15-20 pix in
Photoshop. I use a wacom tablet for sketching.
You need a "script" to pre work each image in the stack, then you modify the corresponding layer with the wacom pencil.
And this is my how to do in Photoshop :
make pics, lowest focal plane first.
The lowest picture is the start, now copy each following picture twice onto the layer stack.
Why twice ? You need one for creating the sharpness mask, the other for adding the corresponding mask.
now create a mask with :
Filter: blur,Smart blur, quality medium, edge only, radius about 48, threshold about 47
Filter: other, maximum, 3-4 pixel
Select: colour range, white pixel with pipette, fuzziness about 25
Select: modify smooth, about 5 pixel
Select: modify border 4 pixel
Now you have a selection to make a mask on the second identical layer.
Add a mask to the second layer of this identical picture.
This mask can now be modified by applying black or white colour to it to control the
visibility of sharp parts of the image.
Thanks,
Gerd
i know Imagepro too, it works well, but price for me is much to high.
I downloaded the trial version of imagepro express 5 today and computed the radiolaria stack
again. And i tested heliconfocus too, the same stack as the pix before.
You can see the results beneath.
Charlie, i usually take 45 minutes on the average to handle about 15-20 pix in
Photoshop. I use a wacom tablet for sketching.
You need a "script" to pre work each image in the stack, then you modify the corresponding layer with the wacom pencil.
And this is my how to do in Photoshop :
make pics, lowest focal plane first.
The lowest picture is the start, now copy each following picture twice onto the layer stack.
Why twice ? You need one for creating the sharpness mask, the other for adding the corresponding mask.
now create a mask with :
Filter: blur,Smart blur, quality medium, edge only, radius about 48, threshold about 47
Filter: other, maximum, 3-4 pixel
Select: colour range, white pixel with pipette, fuzziness about 25
Select: modify smooth, about 5 pixel
Select: modify border 4 pixel
Now you have a selection to make a mask on the second identical layer.
Add a mask to the second layer of this identical picture.
This mask can now be modified by applying black or white colour to it to control the
visibility of sharp parts of the image.
Thanks,
Gerd
I could never manage all that on Photoshop!
The Heliconfocus seems to have fewer artefacts than the IP image, and is a bit cheaper... like 5% of the cost. My IP 4.5 was a cast-off from a research establishment who had a spare copy bought in error. I'd upgrade to 5 but even the upgrade is too pricey for me.
I'd like to see the comparison done with a less indistinct subject. Maybe a histo section? Or an arthropod mount? Or even an opaque macro?
The Heliconfocus seems to have fewer artefacts than the IP image, and is a bit cheaper... like 5% of the cost. My IP 4.5 was a cast-off from a research establishment who had a spare copy bought in error. I'd upgrade to 5 but even the upgrade is too pricey for me.
I'd like to see the comparison done with a less indistinct subject. Maybe a histo section? Or an arthropod mount? Or even an opaque macro?