Description:	BenthicNet is a global compilation of seabed images designed to support the development of automated image processing tools. We currently have the need, but not the resources, to automate the classification of seabed images for the purposes of benthic ecology, mapping, and conservation. BenthicNet provides a volume of images appropriate for the training of machine and deep learning models that may be used for automatic image processing and classification. The dataset is global in scope, containing images from a range of locations, depths, and environments. Images and metadata were obtained from project partners, open access repositories, and third party hosting locations (such as journal supplementary materials), and have been standardized to a common format.
Notes:	Image data within this collection are made available under a variety of terms. The specific license assigned to each image set is listed in the document "00_Documentation/all_licenses_refs.csv" and the full terms and conditions of each license are available in "00_Documentation/licenses".
Authors:	Misiuk, Benjamin; Memorial University of Newfoundland; 0000-0001-5822-4574 Lowe, Scott C.; Dalhousie University; Vector Institute; 0000-0002-5237-3867 Xu, Isaac; Dalhousie University
Keywords:	benthic habitat seabed photograph deep learning computer vision underwater image
Field of Research:	Earth and related environmental sciences > Oceanography > Biological oceanography
Publication Date:	2024-04-30
Publisher:	Federated Research Data Repository / dépôt fédéré de données de recherche
Funder:	Ocean Frontier Institute; Benthic Ecosystem Mapping & Engagement
URI:	https://doi.org/10.20383/103.0614
Related Identifiers:	This dataset is cited by https://github.com/DalhousieAI/benthicnet_probes This dataset is cited by http://arxiv.org/abs/2405.05241
Geographic Coverage:	Place Name Global
Geolocation Box:	West Longitude -180 East Longitude 180 North Latitude 90 South Latitude -90
Appears in Collections:	CIOOS Atlantic / SIOOC Atlantique