Welcome to I3S: Interactive Individual Identification System
We have started beta testing of our new tool: I3
S Pattern! With a grant of
S Pattern is the fourth release in a series of related software tools.
The first release was I3S (now referred to as I3S Classic). This package is based on annotation of the spot centers and is used for animals with lots of similarly shaped spots such as whale sharks. The second package was named I3S Manta or I3S Spot. I3S Manta focuses on animals with less spots and more variation. Instead of annotating the centers only, spots are annotated by the closest fitting ellipse around each spot, in this way taking spot size and shape into account. The third package was I3S Contour, which used the contour of e.g. flukes for identification.
Interested in beta testing? Are you working with turtles of other (marine) animals with natural patterns? Please contact us!
Evaluation of I3S Classic using Whale shark data.
Please tell us your success stories, wishes, bugs etc. so we can do more for you!
Successful conservation activities depend on reliable population information based on distribution, habitat use,
or life history parameters of individual animals. To be able to assess whether a species is endangered,
which conservational measures are most likely to be successful or whether these measures had the expected result,
it is essential to have proper information on existing populations. The population information described above
can only be gathered from many observations of individual animals during a period covering several years.
For recognition of individual animals, often tags attached to the animals are used. Tagging however, is invasive,
costly and tags do not last a lifetime. It also induces stress and when a tag is lost (torn away) it may wound the
animal. Photo-identification using natural marks such as spots and fin or tail shape is proven to be effective over
the entire lifespan, it is relatively cheap and does not induce any stress or damage. The main disadvantage is that
recognition of animals is based on manual inspection of the photo database. As more individuals are added to
photo-identification databases with hundreds or even thousands of individuals, the manual matching process becomes
extremely tedious and therefore looses accuracy. For many marine biologists the efficiency of their research is
seriously obstructed by the absence of tools to assist the identification of individual animals from photos.
Automatic assistance in photo-identification takes away the only disadvantage of this superior technique compared
with tagging. I3S is a free computer-aided photo-identification technique that relies
on natural marks to identify individual animals.
I3S has proven very successful in identification of these species based on
their natural spots. I3S helps the researcher
to extract the spot pattern and then
compares this pattern against all animals in the database and shows the researcher the most relevant results. The
most recent version of I3S is able to recognize 96% of recurring animals in the top 10.
Upon addition to the photo database, I3S expects the researcher to provide information
about the spot pattern. This
step takes about one to two minutes. After this step, the search is entirely automatic and highly accurate.
You can use I3
S free of any charge
. You are allowed to modify it,
copy it, and distribute it under the GNU Public License v2.
For the application you can download
the distribution file
including the source code.