[MARMAM] matching program inquiry update

[Sabre Mahaffy]

Hello all,
I have had an overwhelming response to my inquiry on MARMAM regarding the efficacy and practicality of various cetacean matching programs.  Numerous people have kindly replied with advice and insights into various programs based on personal and second-hand experience, and I have attempted to summarize the responses for all of those interested.  I am certainly not the expert on any of these programs, however as everyone who contributed responses is on MARMAM, if anyone has a question they may be able to chime in and provide answers.
I received responses referencing three matching programs: Finscan, Darwin and I3S* (Interactive Individual Identification System).  The responses were quite varied and in general depend on what you wish to achieve with the matching program.  Finscan and Darwin provide a database component, so for those looking for a database management system for their photo-identification records in addition to a matching aid, these programs would be the ones in which you should focus.  I3S does not provide a database component and is thus a simple ranking system for possible matches.  Note: I only received one response from someone using I3S.
More detailed comments on each program are provided below, but the consensus from researchers who have used Finscan, Darwin or both is that each have their peculiarities and take some getting used to, but that Darwin is the preferable program largely because it is still being technically supported, and is thus able to adapt to the fast-changing computer world.  That said, several of the people who have used one or both programs have indicated that the benefits a matching program provides are most useful in situations where catalogs have not yet been started or are in their infancy, as an established catalog will take time to enter into the program and may not be worth the effort. Many of the researchers who responded were of the opinion that matching by eye was the most efficient and accurate method, but that the database component of a matching program is useful for maintaining sighting records.
The photo software program ACDSee was mentioned as an alternative to the programs mentioned above. Although it does not have an automated matching feature (and thus matching is done manually), it does have features of a database that will allow you to rate and categorize photos, as well as batch-set IPTC and EXIF data. One researcher commented that overall, ACDSee was a faster matching tool than Finscan.
No one provided comments on I3S Contour* specifically so we tested the program over the summer and have summarized our experiences in the table below. We did not conduct a formal trial of the program, so please take our comments accordingly! I3S Contour is similar to Finscan and Darwin in that the fin is traced into the program, attributes about the image are coded, and the fin is then compared against the existing catalog using the matching function. In general we found that the ranking of potential matches was not as accurate as we had hoped, but accuracy improved when more care was taken to trace the entire fin (including a portion of the back) and select photos with as little angle to the camera as possible. I3S does not have a database component and is primarily a quick matching aid.
*http://www.reijns.com/i3s/about/I3S_about.html
Summarized responses to the original MARMAM request for information about the efficacy of available fin matching programs (please note I cannot personally verify or deny any of the information presented on Finscan or Darwin).


Finscan

Darwin

I3S

Technical support

No longer supported

Supported but designed for XP and may not be compatible with Windows 7 or 8

Supported, did not experience any problems with Windows 7

Time investment

~12 min to trace a dorsal into the program (this may be lower with practice)

~2 min to trace a dorsal into the program after practice

~1-2 min to trace a dorsal and add attribute data

Cataloging "other markings"

Does not provide a way to catalog scars (e.g., rake marks), only nicks and notches in the dorsal fin


No response.

Yes, through attribute data, but other markings are not traceable in the program

Cataloging photos/individuals

Database component useful for keeping sighting records and also allows for easy comparison of photos

Database component useful for keeping sighting records and also allows for easy comparison of photos

No database component

Matching aid

Successful match rate is low, even with distinctive individuals. Requires entire fin be visible, exactly the same angle, etc.

Auto-matching feature may not be very practical; accuracy requires photos to be at exactly the same angle. Reviews on this subject are mixed and are likely due to the quality of photos used

Successful match rate is low, even with distinctive individuals. Requires entire fin be visible, exactly the same angle, etc., which improves results.

Usability

Mixed: some found it fairly intuitive without many issues, other did not think it was user friendly

Very user friendly, but can be finicky for individuals that are not highly marked

Very user friendly, simple to install and run

Other program limitations mentioned

No response.

Can only use images from one side, not both (it will flip the photo to create a mirror image)



Feasible for interns

Yes, but expect to do some troubleshooting

Yes, may help reduce matching time by ranking the top matches

Yes, program was largely tested by an intern

Overall efficiency

Faster than matching by eye for large catalogs, but accuracy is highly variable

Not necessarily more efficient than matching by eye, but it might narrow down matches in large catalogs. Best when starting a new catalog instead of converting an existing catalog. Useful for cataloging individuals in a database

Faster than matching by eye for large catalogs, but accuracy is highly variable and checking each individual in the ranked list reduces efficiency


I hope those contemplating using a photo matching program find this summary useful!
Best,
Sabre Mahaffy, M.Sc.
Research Biologist
Cascadia Research Collective
218 1/2 W. 4th Ave.
Olympia, WA
98501 USA
Office 360-943-7325

www.cascadiaresearch.org

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.uvic.ca/pipermail/marmam/attachments/20131126/565f1ce7/attachment.html>