Endless forams: >34,000 modern planktonic foraminiferal images for taxonomic training and automated species recognition using convolutional neural networks
Paleoceanography and Paleoclimatology
By: Allison Y. Hsiang, Anieke Brombacher, Marina Costa Rillo, Maryline J. Mleneck-Vautravers, Stephen Connett, Sian Lordsmith, Anna Jentzen, Michael J. Henehan, Brett Metcalfe, Isabel Fenton, Bridget Wade, Lyndsey Fox, Julie Meilland, Catherine V. Davis, Ulrike Baranowski, Jeroen Groeneveld, Kirsty M. Edgar, Aurore Movellan, Tracy Aze, Harry J. Dowsett, Giles Miller, Nelson Rios, and Pincelli M. Hull
Links
- More Information:
- Open Access Version: Publisher Index Page
- Download citation as: RIS | Dublin Core
Abstract
Accurate planktonic foraminiferal species identification is central to many paleoceanographic studies, from selecting specific species for geochemical research to elucidating the biotic dynamics of microfossil communities relevant to physical oceanographic processes and interconnected phenomena such as climate change. However, species identification varies among taxonomic schools, few resources exist to train students in the difficult task of discerning amongst closely related species, and the number of taxonomic experts is limited. Here, we take the first steps towards removing these rate-limiting steps by generating the first extensive image library of modern planktonic foraminifera, providing taxonomic training tools and resources, and automating species-level taxonomic identification of planktonic foraminifera via machine learning using convolution neural networks. Taxonomic experts identified 34,640 images of modern planktonic foraminifera to the species level. These images are served as species exemplars through the online portal Endless Forams (endlessforams.org) and a taxonomic training portal hosted on the citizen science platform Zooniverse (zooniverse.org/projects/ahsiang/endless-forams/). A supervised machine learning classifier was then trained with more than 24,000 images of planktonic foraminifera and tested using the remaining ~10,000 images (i.e., the validation set). The best classifier provided the correct species name for an image in the validation set 87.4% of the time. Together, these resources provide a rigorous set of training tools in modern planktonic foraminiferal taxonomy and a means of rapidly generating assemblage data via machine learning in future studies.
Suggested Citation
Hsiang, A.Y., Brombacher, A., Costa Rillo, M., Mleneck-Vautravers, M.J., Connett, S., Lordsmith, S., Jentzen, A., Henehan, M.J., Metcalfe, B., Fenton, I., Wade, B., Fox, L., Meilland, J., Davis, C., Baranowski, U., Groeneveld, J., Edgar, K.M., Movellan, A., Aze, T., Dowsett, H.J., Miller, G., Rios, N., Hull, P.M., 2019, Endless forams: >34,000 modern planktonic foraminiferal images for taxonomic training and automated species recognition using convolutional neural networks: Paleoceanography and Paleoclimatology, v. 34, p. 1157-1177, https://doi.org/10.1029/2019PA003612.
| Publication type | Article |
|---|---|
| Publication Subtype | Journal Article |
| Title | Endless forams: >34,000 modern planktonic foraminiferal images for taxonomic training and automated species recognition using convolutional neural networks |
| Series title | Paleoceanography and Paleoclimatology |
| DOI | 10.1029/2019PA003612 |
| Volume | 34 |
| Publication Date | July 22, 2019 |
| Year Published | 2019 |
| Language | English |
| Publisher | American Geophysical Union |
| Contributing office(s) | Eastern Geology and Paleoclimate Science Center |
| Description | 21 p. |
| First page | 1157 |
| Last page | 1177 |