@article{fdi:010089628,
  title = {{T}oward an artificial intelligence-assisted counting of sharks on baited video},
  author = {{V}illon, {S}{\'e}bastien and {I}ovan, {C}orina and {M}angeas, {M}organ and {V}igliola, {L}aurent},
  editor = {},
  language = {{ENG}},
  abstract = {{G}iven the global biodiversity crisis, there is an urgent need for new tools to monitor populations of endangered marine megafauna, like sharks. {T}o this end, {B}aited {R}emote {U}nderwater {V}ideo {S}tations ({BRUVS}) stand as the most effective tools for estimating shark abundance, measured using the {M}ax{N} metric. {H}owever, a bottleneck exists in manually computing {M}ax{N} from extensive {BRUVS} video data. {A}lthough artificial intelligence methods are capable of solving this problem, their effectiveness is tested using {AI} metrics such as the {F}-measure, rather than ecologically informative metrics employed by ecologists, such as {M}ax{N}. {I}n this study, we present both an automated and a semi-automated deep learning approach designed to produce the {M}ax{N} abundance metric for three distinct reef shark species: the grey reef shark ({C}archarhinus amblyrhynchos), the blacktip reef shark ({C}. melanopterus), and the whitetip reef shark ({T}riaenodon obesus). {O}ur approach was applied to one-hour baited underwater videos recorded in {N}ew {C}aledonia ({S}outh {P}acific). {O}ur fully automated model achieved {F}-measures of 0.85, 0.43, and 0.72 for the respective three species. {I}t also generated {M}ax{N} abundance values that showed a high correlation with manually derived data for {C}. amblyrhynchos ({R} = 0.88). {F}or the two other species, correlations were significant but weak ({R} = 0.35-0.44). {O}ur semi-automated method significantly enhanced {F}measures to 0.97, 0.86, and 0.82, resulting in high-quality {M}ax{N} abundance estimations while drastically reducing the video processing time. {T}o our knowledge, we are the first to estimate {M}ax{N} with a deep-learning approach. {I}n our discussion, we explore the implications of this novel tool and underscore its potential to produce innovative metrics for estimating fish abundance in videos, thereby addressing current limitations and paving the way for comprehensive ecological assessments.},
  keywords = {{D}eep learning ; {N}eural network ; {C}oral reef ; {M}arine ecology ; {S}hark conservation},
  booktitle = {},
  journal = {{E}cological {I}nformatics},
  volume = {80},
  numero = {},
  pages = {102499 [9 p.]},
  ISSN = {1574-9541},
  year = {2024},
  DOI = {10.1016/j.ecoinf.2024.102499},
  URL = {https://www.documentation.ird.fr/hor/fdi:010089628},
}