The individual identification of the sharks of a population is the fundamental step to be taken if we want to conduct studies such as population estimates, behavioral evolution with age, hierarchy identification. When working with large and open populations, this process becomes even more difficult. One of the most commonly used methods is the identification of individuals through the photograph of their dorsal fin, which can act as a fingerprint for considerable time intervals. Comparing images is a time-con- suming task, for human operators; however, defining automatic systems for image comparisons is not easy, because of variations in lighting, saturation, contrast, orientation and position of objects in images. In this work we propose an alternative approach for helping at automatically identifying sharks from pictures of their fins, by taking advantage from results in different areas of computer science. Indeed, sequence alignment is a common task in Computer Science: two sequences are compared in order to understand whether they feature some common areas. For instance, this is extensively used in bioinformatics in order to identify regions of similarity between DNA sequences. In the proposed approach, images of fins are not directly compared; instead, using advanced image processing techniques, contour of fins are recognized, extracted and then represented as ordered sequences of symbols. Each sequence representing a fin is stored in a database along with the identifier of the associated shark, so that, given a picture of a fin, various sequences alignment and machine learning techniques are employed in order to identify the most cor- related fin stored in the database. Eventually, this allows users to uniquely identify a shark from new unseen photos of it.

Individual identification of sharks through dorsal fin: a new approach thanks to advances in computer science. Poster Abstract.

Gianni Giglio;Aldo Marzullo;Francesco Calimeri;Francesco Cauteruccio;Giovambattista Ianni;Chiara Romano;Emilio Sperone;Sandro Tripepi;Giorgio Terracina
2018-01-01

Abstract

The individual identification of the sharks of a population is the fundamental step to be taken if we want to conduct studies such as population estimates, behavioral evolution with age, hierarchy identification. When working with large and open populations, this process becomes even more difficult. One of the most commonly used methods is the identification of individuals through the photograph of their dorsal fin, which can act as a fingerprint for considerable time intervals. Comparing images is a time-con- suming task, for human operators; however, defining automatic systems for image comparisons is not easy, because of variations in lighting, saturation, contrast, orientation and position of objects in images. In this work we propose an alternative approach for helping at automatically identifying sharks from pictures of their fins, by taking advantage from results in different areas of computer science. Indeed, sequence alignment is a common task in Computer Science: two sequences are compared in order to understand whether they feature some common areas. For instance, this is extensively used in bioinformatics in order to identify regions of similarity between DNA sequences. In the proposed approach, images of fins are not directly compared; instead, using advanced image processing techniques, contour of fins are recognized, extracted and then represented as ordered sequences of symbols. Each sequence representing a fin is stored in a database along with the identifier of the associated shark, so that, given a picture of a fin, various sequences alignment and machine learning techniques are employed in order to identify the most cor- related fin stored in the database. Eventually, this allows users to uniquely identify a shark from new unseen photos of it.
2018
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.11770/340404
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact