July 2, 2013 Featured Article Read More →

Can photographs be used to “tag” fish?

By David Bradley.

In order to successfully conserve a species, information must be gathered on its distribution, habitat and the dynamics of its population, such as the population size. This information can then be used to assess whether or not a species is endangered, which management strategies have the best chance of working and, once deployed, which management strategies appear to be working.

To gather this information long term observational studies of individuals must be carried out and in order to do this, individual animals have to be marked so they can be identified upon recapture. This is known as mark and recapture. One method of mark and recapture which is frequently employed is the use of a variety of tags. This method however is not entirely fool proof, as maintaining tagging programmes and data bases can be expensive and the tags may not last throughout the entire life of the animal creating bias.

image

Recently, with the advent of cheap, high quality cameras, researchers have started to see potential in the use of photographic identification of individuals. This has many benefits over traditional methods. For example, databases can be built up with minimal effort. The major disadvantage of this however, is that most of the time individual photographs must be compared by eye. As the database grows in size to hundreds, perhaps even thousands of images, this can become tedious and laborious which can have a negative impact on accuracy.

To get around this problem, developers have come up with several computer software packages to analyse the photographs and match them to known individuals. One such programme is Interactive Individual Identification System, or I3S.

I3S is a free programme which assists users in identifying individuals from photographs. It was originally developed for use with ragged toothed sharks, but has recently been applied to several other species of elasmobranch including manta rays. It works by allowing researchers to mark the outline of patterns on an unknown individual. This outline is then stored as a “digital fingerprint” unique to that individual. This “digital fingerprint” can then be compared to a database of known fish. Once compared, the software then ranks known images in order of the most likely matches. It is then up to users to compare the images by eye and determine whether or not there are conclusive matches. Recently researchers at The University of Glasgow have tested its effectiveness at recognising common skate individuals.

D_Bradley

The above image shows a photograph of a common skate after it has been “digitised” using I3S (left) and the “digital fingerprint” of the same fish (right) that is used to make comparisons.

To test whether this technique was applicable to common skate, researchers asked anglers to send in photographs of their hard earned catches. These photographs were then added to the database and known photographs were compared with the database. Two areas of the fish were assessed, the wing and the dorsal area.

It was found that patterns on the wing gave low levels of accuracy in terms of matching the photograph to its counterpart, the most likely explanation of this is that the pattern found on the wing was not unique enough to the individual to allow for comparison. The results from the dorsal comparisons told a different story however, with the correct individual being ranked first 83% of the time and ranked in the top ten 94% of the time.

Although this result shows that the technique has promise in use with common skate there are still several questions which need to be answered before it can be fully implemented. For example, it is unknown whether or not the markings change throughout the life of a skate which may hinder photographic identification over longer periods of time. As such, longer term studies need to be conducted.

In the meantime, anglers wanting to take photographs of fish to aid future research can help by bearing in mind these guidelines when taking pictures:

· Photographs should be taken on a flat surface, at right angles to the fish

· If possible, keep the flash off to reduce glare

· Finally, feel free to take a few pictures, as different pictures of the same fish help with accuracy

If anybody would like further information on the project, don’t hesitate to contact the author by using the “contact us” facility at the top of this page and stating in your message that you would like to contact David Bradley.

Posted in: Shark Bites
Rss Feed Facebook button Webonews button Digg button Flickr button Newsvine button