Jump to content


"Technical" Photos


  • This topic is locked This topic is locked
25 replies to this topic

#1 Guest_gitano_*

Guest_gitano_*
  • Guests

Posted 15 January 2014 - 01:20 PM

I'm new to NANFA, so it is certainly possible that I have not been able to find an existing tutorial on photographing fish. Given that, if there is such a tutorial, might I suggest that members (and non-members alike if they are allowed to post images), be encouraged to put a scalar of some sort in their images. I appreciate the "artist nature" of some of the pictures, but there is enormous value in the morphometric data that can be extracted from these images. A simple scalar allows that data to be collected and recorded in perpetuity. It is a shame to have a fish 'in hand' and 'record the moment' and not simply add something to the image of known size that allows morphometric analysis of the image. In deference to the artistic nature of things, the scalar can be ANYTHING of KNOWN dimension. Being a scientist, my photarium has "ugly" ruling as the background. However, if a photarium is being used and the photographer has an "artistic" background, simply noting the dimension of one of the items in the background is all that is required. For example: if there was a blade of Vallisneria sp. in the background, knowing the width of that blade would be sufficient information to allow the collection of morphometric data from that image.

Here's an example of a couple of typical "technical" images collected for use by Biopar (www.biopar.com).

The first picture - Onchorhynchus mykiss - is a "field" acquired image. You will note the vertical lines spaced 1" apart.

The second picture is a hatchery acquired image of O. tshawytschia. (The vial contains a genetic sample.)

The third picture is the automatically derived mid-eye-to-fork (MEF), and fork (FL), lengths.

These images were acquired for research purposes and are anything but "artistc", but they illustrate the "tip of the iceberg" of the power of morphometric analysis.

Paul

Edited by gitano, 15 January 2014 - 01:24 PM.


#2 Guest_gerald_*

Guest_gerald_*
  • Guests

Posted 15 January 2014 - 02:06 PM

Hadn't heard of "mid-eye-to-fork length" until now. Is that something unique to salmon biology because of the funky snout growth? Welcome, and thanks for the good ideas. There are several threads on photo-tanks ("photarium" is another new term to me) and fish photography methods if you search around the forum here.

#3 Guest_gitano_*

Guest_gitano_*
  • Guests

Posted 15 January 2014 - 02:33 PM

Hello Gerald,

You are correct that MEF is a response to skewed length data caused by non-linear growth of the rostrum (nose) of spawing salmonids. (It's called a "kype".) Especially males. It is also appropriate for hatchery fish that tend to blunt their noses on raceway and tank walls.

I have seen various threads that include advice on taking pictures, but no specific NANFA tutorial on the subject.

Paul

#4 Guest_daveneely_*

Guest_daveneely_*
  • Guests

Posted 15 January 2014 - 02:52 PM

For example: if there was a blade of Vallisneria sp. in the background, knowing the width of that blade would be sufficient information to allow the collection of morphometric data from that image.


Hi Paul,

Thanks for the suggestion. If you poke around, you might find some images with scale bars...

I have to say, though -- if you're relying upon the width of a Vallisneria blade that's not the same distance away from the camera lens as the fish as a reference... well, let's just say I'd be somewhat suspect of that data set.

Mark Sabaj (of the Academy of Sciences at Philadelphia) and others put together a pretty slick tutorial on fish specimen photography. It's worth tracking down.

Oh, and I almost forgot... Welcome!

Dave

#5 Guest_daveneely_*

Guest_daveneely_*
  • Guests

Posted 15 January 2014 - 02:59 PM

Now wait just one second... I really hope you're not data mining NANFA's Photo Archive to use to promote your own proprietary algorithm and/or software.

I'm not sure I'd be down with that.

#6 Guest_gitano_*

Guest_gitano_*
  • Guests

Posted 15 January 2014 - 03:13 PM

It is certainly true that the scalar - whatever it is - must be "close" to the same plane as the fish in question in order for size estimates to have a reasonable precision. However, in a photarium, the background is almost always closer than necessary for a precision greater than can be obtained with physical measurements. In a controlled environment like a photarium, precision of 0.01 mm is easily attained. I would report the level of precision possible with today's digital cameras, but people would get hung up on the trivia instead of focusing on the point which is including a scalar of known dimension.

I'm not personally looking for "how" to take fish pictures as I have been doing that for more than 50 years. I am suggesting that NANFA, and specifically the NANFA forum, PROVIDE a tutorial. Within that tutorial it would be "good" to demonstrate the value of some form of scalar in the image, AND, as you point out, emphasize the need for that scalar to be as close to the plane of the subject as possible. But it should also be pointed out that excellent precision can still be obtained when the scalar is NOT in the EXACT plane of the fish.

Paul

#7 Guest_gitano_*

Guest_gitano_*
  • Guests

Posted 15 January 2014 - 03:18 PM

Please read my post "New NANFA Member" in the "Welcome to the NANFA Forum" site. I would copy the link to the thread but apparently I can't do that until my membership has been "approved".

Paul

Edited by gitano, 15 January 2014 - 03:44 PM.


#8 Guest_Erica Lyons_*

Guest_Erica Lyons_*
  • Guests

Posted 15 January 2014 - 03:23 PM

Please read my post "New NANFA Member" in the "Welcome to the NANFA Forum" site. I would copy the link to the thread but apparently I can't do that until my membership has been "approved". Not too excited about NANFA taking money "instantly" but having to wait days for "approval". Not really a "big" deal. Just a bit gauche.

Paul

Oh, there used to be spam robots that would make a username and then immediately after joining, post a dozen hyperlinks to buy knockoff louis vitton shoes.

To be helpful, here's the link. http://forum.nanfa.o...w-nanfa-member/

#9 Guest_gitano_*

Guest_gitano_*
  • Guests

Posted 15 January 2014 - 04:06 PM

Thanks, Erica. Like I said, not too big a deal. But not being a member limits access and use. For the record, I just received notice from Tom Watson welcoming me to NANFA, so I assume I have been "approved". However, I am still showing "Guest" in the ID icon.

I run two "forums", one "professional" for Biopar,com, and another "private" one, so I fully understand the morass of ne'er-do-wells "out there".

Thanks again,
Paul

#10 Guest_natureman187_*

Guest_natureman187_*
  • Guests

Posted 17 January 2014 - 08:06 PM

I have seen various threads that include advice on taking pictures, but no specific NANFA tutorial on the subject.


It takes a NANFA member to create one. Seeing you are a member, make one for us!

#11 Guest_gitano_*

Guest_gitano_*
  • Guests

Posted 17 January 2014 - 08:50 PM

I'm having serious trouble with the editor. First, after a few characters, it slows down to the point that is it impossible to use. I have to type each character 5 or more times to get it to show up. To work around that I thought I'd compose the text in Word and "cut and paste" here. Cannot paste ANYTHING. Is there some secret handshake required to "paste" to this editor. I can paste NOTHING, no links, no text, nada.

I have attached my response to the topic as an attached image.

Paul

#12 Michael Wolfe

Michael Wolfe
  • Board of Directors
  • North Georgia, Oconee River Drainage

Posted 17 January 2014 - 09:05 PM

Paul,

I think you misunderstand who NANFA really is. It is an all volunteer organization. There are not any "powers that be".

Anything that gets done around here is started by someone who is interested. It will not get done by suggesting that someone else do it. It will get done by starting, and then (if as you suggest above, you want to do) reaching out to others and engaging them. Create something wonderful and the world will beat a path to your door.

And if you continue to have problems with the browser, please let me know... start a thread in the "bugs..." section.
Either write something worth reading or do something worth writing. - Benjamin Franklin

#13 Guest_gitano_*

Guest_gitano_*
  • Guests

Posted 17 January 2014 - 09:19 PM

If NANFA is as you describe it, and you should know, then indeed I do misunderstand NANFA.

That said, there are indeed "powers that be". SOMEONE is responsible for this website. Someone is responsible for organizing the annual meetings. And so forth. That is the "powers that be" to which I referred.

With respect to the editor, I am still having a problem, and will start a thread in the "bugs" section. For the moment, I have one diagnostic for you: The problem starts as soon as I use the "Enter" key. In other words, when I start a new paragraph.

Below is a sentence typed without 'correcting' what the editor does:
Te quk brn f jpd ovtheazy d. It should read "The quick brown fox jumped over the lazy dog."

Still can'taste

Paul

#14 Guest_gitano_*

Guest_gitano_*
  • Guests

Posted 17 January 2014 - 09:39 PM

I tried to edit the above post to add a post script, but could not. The first time I tried to add a new post to add the PS, the editor told me repeatedly that the post was "empty". Now it's "acting up" even without using the "Enter" key.

I note that NANFA has Board of Directors, an Annual Convention, gives grants, etc. and there are no "powers that be"?

Paul

#15 Michael Wolfe

Michael Wolfe
  • Board of Directors
  • North Georgia, Oconee River Drainage

Posted 17 January 2014 - 09:46 PM

Users are not allowed to "edit posts" after a very short period of time (a forum rule that was put in place a long time ago).

Yes, I am on the board of directors. Yes, our conventions are organized by a member volunteer. Yes, we have volunteer grant committees.

Don't get defensive, I am just suggesting that as a new member, that you jump in and start something if you want to see it happen. If it gets traction here with the membership then you will have something.
Either write something worth reading or do something worth writing. - Benjamin Franklin

#16 Guest_gitano_*

Guest_gitano_*
  • Guests

Posted 18 January 2014 - 12:58 PM

I'm not getting defensive. NANFA just appears to be a "well-oiled machine" with a very distinct organization. That doesn't happen without some "adult supervision". I own and run two internet forums, and fully "appreciate" newbies showing up and telling me what I "should do". As I agreed, I did indeed misinterpret what NANFA is. Any "terseness" on my part is due to the editing issue, and should not be interpreted as defensiveness.

I WILL (editing tools turned off in order to facilitate typing, so no means of adding emphasis other than ALL CAPS), start A - not THE - ball rolling with a thread about "technical fish pictures". We can have an off-line conversation about what and how I think that thread should be administered. If we can come to a meeting of the minds about that, I will 'get on it'.

Thanks,
Paul

#17 Guest_gitano_*

Guest_gitano_*
  • Guests

Posted 24 February 2014 - 03:15 PM

While it doesn't appear that the NANFA forum is a good 'fit' for me, I don't like to leave things I start unfinished. Given that, and in the context of this thread, I am providing some images and data to illustrate the use of TNT and how I think it might have been of some use to the members of NANFA.

In January of this year I received some juvenile fish from a NANFA member. (I have not requested permission from them to publish their name, so I won't. They may, after reading this thread, choose to make their contribution public.) I ordered 3 pairs of young-of-the-year fish: Two each of Redbreasted (Lepomis auritus), pumpkinseed (L. gibbosus), and dollar (L. marginatus). At the time of receipt, they were all between 30 and 40 mm long. I photographed three of them, but due to the stress of shipment I decided to put off further photography until they had demonstrated willingness to eat and had a chance to acclimate to their new environment. On January 31 - approximately 15 days after receipt - I performed a water change and chose that time to acquire the first TNT-appropriate pictures. Here are some representatives of those first pictures, with the TNT annotations.

The first image is of a fish with each point of morphometric interest (17 of them) as well as four lines, in this case the longest soft rays of the dorsal, anal, pelvic and pectoral fins.
IMG_0350.JPG duplicated.jpg Snapshot.jpg
The second image is a table of the extracted data from each point and the lengths of the lines.
350 data.JPG

The third image is of the same fish at the next water change two weeks later, and it's associated data table.
IMG_6674.JPG rotated -7.86 duplicated.jpg Snapshot.jpg
6674 data.JPG
You will note a couple of changes in the data. I have included the length of the longest spine in the dorsal fin, and rearranged two of the morphometric point numbers. I have also registered the rest of the image to point #1. This facilitates post acquisition computation and data processing. Modifications such as this are often the case when starting a new species or project. It may take three or four iterations to determine the best format of data to collect.

While this is just one side of one of the six fish, all fish were photographed on both sides, and data recorded for image acquired. Subsequent to the February 15 photo session, I was able to look at - among other things - growth and growth rate for each fish. The next two images are the table of that data and the graphic representation of that data.
1st 2wk growth data tbl.JPG
(Within fish side-to-side length measurements are all within 0.5mm.)

1st 2wk growth data graph.JPG

The fish were kept at a temperature of 23C, plus or minus 1C, and fed twice daily to satiation. The average growth for the 15-day interval was 7.2mm and the average per day growth was 0.48mm.

This is just a "quick and dirty" illustration of the potential uses of TNT. It could be as scientifically rigorous or lax as one desires. My thought for its most valuable use to NANFA would have been to "tag" all of the fish in a given body of water that is "regularly" sampled by a member, or by those collectively and systematically sampling a specific body of water. In a relatively short interval of time, one would "know" essentially all of the fish in the body of water, and any new fish captured would most likely be recruits. Clearly, growth rates could be recorded not only for individuals but also for populations, and comparisons made not just between populations, but also within populations after a stressor event or after some "restoration" effort. Morphometric variances between populations, genders, breeding condition, etc are trivially quantified. Add to the metadata gps information and some very interesting analysis could be undertaken.

NANFA has "at its disposal" a small army of interested individuals that, collectively, could assemble an unprecedented database of rigorously quantified data. The ONLY requirement being to take pictures of the fish 'you' already have in your hands. Speaking of "in hand", watching these fish of mine from "day one" will allow provide information on these three species that will allow identification of species at an earlier age than is currently possible. For those raising fish, starting the process as early as possible in the life history would allow the compilation of a "key" that would facilitate very early life history identification of wild fish. When people work toward their own interests, but share data, the potential is almost limitless.

So... for those individuals that would like to set up FREE accounts to start collecting and maintaining data on your captive fish as well as wild fish you may sample, send me a PM and I will set up a FREE account for you at Biopar. For those of you that continue believe that this is some sort of business effort on my part, the nicest thing I can say to you is that we are usually our own worst punishment.

Paul

#18 Guest_Erica Lyons_*

Guest_Erica Lyons_*
  • Guests

Posted 24 February 2014 - 03:30 PM

Is there a way to use this to quantify a photo and be able to definitively say whether a fish is elassoma evergladei versus its okefenokee and gilberti genus mates?

Evergladei:
Posted Image
Posted Image
http://gallery.nanfa... right.jpg.html
http://gallery.nanfa...g left.jpg.html

Gilberti:
Posted Image
Posted Image
Posted Image
http://gallery.nanfa...l crop.jpg.html
http://gallery.nanfa...ze_002.jpg.html
http://gallery.nanfa...unfish.jpg.html

#19 Guest_gitano_*

Guest_gitano_*
  • Guests

Posted 24 February 2014 - 06:50 PM

Hello Erica - The quality of those images are "fine" for individual ID ("tagging"). With respect to discriminating between species and sub-species, the "quality" of hte above images isn't the issue, it is 1) lack of a scalar in the image (except for the last one), 2) and knowledge of the taxonomic discriminants. Even if color is one of the discriminants, a "scalar" (color card) in the image would be best. In the case of wave-length pattern ID (color), the "white balance" used, relative to the actual light source (incandescent, fluorescent, "natural", etc.) would have an impact on the accuracy of the analysis. We can do some very sophisticated spectral analysis, but without a priori knowledge of the differences between classes (species, sub-species, etc.), we couldn't say with certainty that one fish was "one thing" and another different. The goes for any taxonomic differences in morphometry.

These images would be fine for individual ID ("tagging"). These fish could be known for their life from these images.

The big difference that quality of image makes is in commercial applications, and the difference is as you might assume, cost. For people setting up a "recreational" account that might not input more than say 25 fish at the most at one time, processing the images (extracting the information) can be accomplished "by hand". However, we recently photographed part of a cohort of Oncorhynchus mykiss at a hatchery prior to its release. We photographed 12,000 fish of that cohort as a means of non-invasive "tagging". All of their pictures were taken in two days, and we were able to rigorously control the image acquisition. Depending on the quality of the image, processing "by hand" requires between 5 and 20 minutes per image. Automatic processing takes about 20 msec per image. You can imagine the labor costs if it had been necessary to process 12,000 images "by hand".

Recognizing - through experience - the differences between two classes (species, genders, races, etc) would allow one to use TNT to establish some quantitative differences. For example, I am measuring the lengths of the fins to see if there is 1) differential growth as a function of life-stage, 2) differences between species in the ratios of fin-length to total length, and 3) to see if there are differences in gender within species. If these differences show up this early in these species life history, then not only may we be able to differentiate specie earlier than every before, we might be able to differentiate gender at a very early age. In the situation where a breeder KNOWS the species because they segregate their fish by species, it could be extremely valuable to know the genders of the YOYs both for commercial reasons (consider the differences in cost of chicken chicks between "straight run" (gender unknown), and "cockerels" vs. "pullets"), and for looking at environmental factors (like incubation temperature), that might effect gender ratios within a cohort.

Paul

PS - There are circumstances in which a scalar in the image wouldn't be necessary. Let's say one wanted to look at the images of two fish to determine if they were different species AND one knew that the "nose length" of species "A" was 11% of total length, and the "nose length" of species "B" was 15%. Proportional measurements require no scalar other than the length of the fish or whatever particular metric one was using.

Paul

#20 Guest_Erica Lyons_*

Guest_Erica Lyons_*
  • Guests

Posted 24 February 2014 - 10:13 PM

There are circumstances in which a scalar in the image wouldn't be necessary. Let's say one wanted to look at the images of two fish to determine if they were different species AND one knew that the "nose length" of species "A" was 11% of total length, and the "nose length" of species "B" was 15%. Proportional measurements require no scalar other than the length of the fish or whatever particular metric one was using.

I think those sorts of proportional measurements are known for evergladei and okefenokee/gilberti. Here is a scientific paper: http://biology.unm.e...et al. 2009.pdf I can't copy and paste the millimeter descriptions from it but I can link to it.
Evergladei: http://fishbase.sini...ary.php?ID=3364

Not being able to tell the two apart is a huge problem for vendors. If there was a program that they could enter a picture in and, based on the ratio of fin size to body, nose length, fin ray count, other ratios, etc tell them what species it is, that would be extremely valuable. Not just for elassoma, but for basically all the fish we find in the field and need to identify. Notropis, Lepomis, everything would really benefit from rapid computer analysis and ID.

Can we as NANFA add that as a tool to our website? Take a picture of your fish in the field with a scale bar in the ID tank and have an ID result returned immediately? If we bought a subscription as a group, we could add it as a feature for members. Imagine being able to within seconds return the result of exactly which species of shiner that was. That would be super useful. Lots of people would join NANFA just for the ability to use that kind of tool.

Paul, would your program be able to return species ID based on pictures taken with scale bars and the known ratios of fin lengths, eye lengths, etc of the species?




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users