Journal Article
Machine learning techniques to characterize functional traits of plankton from image data
Eric C. Orenstein; Sakina‐Dorothée Ayata; Frédéric Maps; Érica C. Becker; Fabio Benedetti; Tristan Biard; Thibault de Garidel‐Thoron; Jeffrey S. Ellen; Filippo Ferrario; Sarah L. C. Giering; Tamar Guy‐Haim; Laura Hoebeke; Morten Hvitfeldt Iversen; Thomas Kiørboe; Jean‐François Lalonde; Arancha Lana; Martin Laviale; Fabien Lombard; Tom Lorimer; Séverine Martini; Albin Meyer; Klas Ove Möller; Barbara Niehoff; Mark D. Ohman; Cédric Pradalier; Jean‐Baptiste Romagnan; Simon‐Martin Schröder; Virginie Sonnet; Heidi M. Sosik; Lars S. Stemmann; Michiel Stock; Tuba Terbiyik‐Kurt; Nerea Valcárcel‐Pérez; Laure Vilgrain; Guillaume Wacquet; Anya M. Waite; Jean‐Olivier Irisson
Limnology and Oceanography · Vol. 67, Issue 8 · pp. 1647-1669 · 2022
Abstract
Plankton imaging systems supported by automated classification and analysis have improved ecologists' ability to observe aquatic ecosystems. Today, we are on the cusp of reliably tracking plankton populations with a suite of lab‐based and in situ tools, collecting imaging data at unprecedentedly fine spatial and temporal scales. But these data have potential well beyond examining the abundances of different taxa; the individual images themselves contain a wealth of information on functional traits. Here, we outline traits that could be measured from image data, suggest machine learning and computer vision approaches to extract functional trait information from the images, and discuss promising avenues for novel studies. The approaches we discuss are data agnostic and are broadly applicable to imagery of other aquatic or terrestrial organisms.