26,000+ machine-identified images of Modern planktonic foraminifera from 23 Atlantic coretops

• Main Authors: Hsiang, Allison Y., Hull, Pincelli M.
• Format: Other/Unknown Material
• Language: unknown
• Published: Zenodo 2022
• Subjects: Planktonic foraminifera; AutoMorph; Deep learning classification; Microfossil; Endless Forams; Fenton; Metcalfe
• Online Access: https://doi.org/10.5281/zenodo.7268451

This dataset contains 26,663 images of Modern planktonic foraminifera taken as part of the Endless Forams (endlessforams.org) initiative that did not receive species labels assigned by human classifiers in Hsiang et al. (2019). These images were processed using AutoMorph (Hsiang et al. 2018) and assigned species labels using a deep learning classifier with a validation accuracy of 87.4% (Hsiang et al. 2019). The machine-assigned species labels for these images are available as part of the supplementary data associated with Hsiang & Hull (in prep., under screening at bioRxiv). The code associated with the generation of this dataset is available at https://github.com/palaeomachinist/foram-comm-ecol. References Hsiang AY, Brombacher A, Rilo MC, Mleneck-Vautravers MJ, Conn S, Lordsmith S, Jentzen A, Henehan MJ, Metcalfe B, Fenton I, Wade B, Fox L, Meilland J, Davis CV, Baranowski U, Groeneveld J, Edgar KM, Movellan A, Aze T, Dowsett H, Miller G, Rios N & Hull PM (2019) Endless Forams: >34,000 modern planktonic foraminiferal images for taxonomic training and automated species recognition using convolutional neural networks. Paleoceanography and Paleoclimatology . 34(7):1157-1177. ( https://doi.org/10.1029/2019PA003612 ) Hsiang AY, Nelson K, Elder LE, Sibert EC, Kahanamoku SS, Burke JE, Kelly A, Liu Y, and Hull PM (2018) AutoMorph : Accelerating morphometrics with automated 2D and 3D image processing and shape extraction. Methods in Ecology and Evolution. 9(3):605-612. ( https://doi.org/10.1111/2041-210X.12915) Hsiang AY & Hull PM. Automated community ecology using deep learning: a case study of planktonic foraminifera. AYH was supported by Vetenskapsrådet (Swedish Research Council) grant 2020-03515.