Wild Cetacean Identification using Image Metadata

POLLICELLI DÉBORA; COSCARELLA MARIANO; DELRIEUX CLAUDIO

San Luis

Congreso; XXII Congreso Argentino de Ciencias de la Computación 2016; 2016

Red UNCI

Identification of individuals in marine species, especially in Cetacea,is a critical task in several biological and ecological endeavours. Most of thetimes this is performed through human-assisted matching within a set ofpictures taken in different campaigns during several years and spread aroundwide geographical regions. This requires that the scientists perform laborioustasks in searching through archives of images, demanding a significantcognitive burden which may be prone to intra and inter observer operationalerrors. On the other hand, additional available information, in particular themetadata associated to every image, is not fully taken advantage of. The presentwork presents the result of applying machine learning techniques over themetadata of archives of images as an aid in the process of manual identification.The method was tested on a database containing several pictures of 230 different Commerson dolphins (Cephalorhynchus commersoni) taken over aspan of seven years. A supervised classifier trained with identifications made bythe researchers was able to identify correctly above 90% of the individuals onthe test set using only the metadata present in the image files. This reducessignificantly the number of images to be manually compared, and therefore thetime and errors associated with the assisted identification process.