A realistic fish-habitat dataset to evaluate algorithms for underwater visual analysis.


Journal

Scientific reports
ISSN: 2045-2322
Titre abrégé: Sci Rep
Pays: England
ID NLM: 101563288

Informations de publication

Date de publication:
04 09 2020
Historique:
received: 18 11 2019
accepted: 14 08 2020
entrez: 5 9 2020
pubmed: 6 9 2020
medline: 20 3 2021
Statut: epublish

Résumé

Visual analysis of complex fish habitats is an important step towards sustainable fisheries for human consumption and environmental protection. Deep Learning methods have shown great promise for scene analysis when trained on large-scale datasets. However, current datasets for fish analysis tend to focus on the classification task within constrained, plain environments which do not capture the complexity of underwater fish habitats. To address this limitation, we present DeepFish as a benchmark suite with a large-scale dataset to train and test methods for several computer vision tasks. The dataset consists of approximately 40 thousand images collected underwater from 20 habitats in the marine-environments of tropical Australia. The dataset originally contained only classification labels. Thus, we collected point-level and segmentation labels to have a more comprehensive fish analysis benchmark. These labels enable models to learn to automatically monitor fish count, identify their locations, and estimate their sizes. Our experiments provide an in-depth analysis of the dataset characteristics, and the performance evaluation of several state-of-the-art approaches based on our benchmark. Although models pre-trained on ImageNet have successfully performed on this benchmark, there is still room for improvement. Therefore, this benchmark serves as a testbed to motivate further development in this challenging domain of underwater computer vision.

Identifiants

pubmed: 32887922
doi: 10.1038/s41598-020-71639-x
pii: 10.1038/s41598-020-71639-x
pmc: PMC7473859
doi:

Types de publication

Journal Article Research Support, Non-U.S. Gov't

Langues

eng

Sous-ensembles de citation

IM

Pagination

14671

Références

IEEE Trans Pattern Anal Mach Intell. 2017 Apr;39(4):640-651
pubmed: 27244717
Foods. 2017 Jul 19;6(7):
pubmed: 28753923

Auteurs

Alzayat Saleh (A)

James Cook University, Townsville, Australia. alzayat.saleh@my.jcu.edu.au.

Issam H Laradji (IH)

University of British Columbia, Vancouver, Canada.
Element AI, Montreal, Canada.

Dmitry A Konovalov (DA)

James Cook University, Townsville, Australia.

Michael Bradley (M)

James Cook University, Townsville, Australia.

David Vazquez (D)

Element AI, Montreal, Canada.

Marcus Sheaves (M)

James Cook University, Townsville, Australia.

Articles similaires

Robotic Surgical Procedures Animals Humans Telemedicine Models, Animal

Odour generalisation and detection dog training.

Lyn Caldicott, Thomas W Pike, Helen E Zulch et al.
1.00
Animals Odorants Dogs Generalization, Psychological Smell
Animals TOR Serine-Threonine Kinases Colorectal Neoplasms Colitis Mice
Animals Tail Swine Behavior, Animal Animal Husbandry

Classifications MeSH