An Image Dataset of Cleared, X-Rayed, and Fossil Leaves Vetted to Plant Family for Human and Machine Learning.

WILF PETER; SCOTT WING; HERBET W. MEYER; JACOB ROSE; ROHIT S.; SERRE THOMAS; RUBÉN CÚNEO; DONOVAN, MICHAEL P.; ERWIN DIANE; GANDOLFO M. ALEJANDRA; GONZALEZ-AKRE ERIKA; FABIANY HERRERA; HU SHUSHENG; ARI IGLESIAS

Virtual

Conferencia; Botanical Society of America Botany Conference; 2021

Botanical Society of America

Leaves are the most abundant and visible plant organ, both in the modern world and the fossil record. Assigningfoliage to the correct plant family based on leaf architecture is a fundamental botanical skill that is critical foridentifying isolated fossil leaves, which often, especially in the Cenozoic, represent extinct genera and species fromextant families. Resources focused on leaf identification are remarkably scarce; however, the situation hasimproved due to the recent proliferation of digitized herbarium material, live-plant image libraries and identificationapps, and online collections of cleared and fossil leaf images. Nevertheless, the need remains for a specializedimage dataset for comparative leaf architecture. We address this need here by assembling an open-accessdatabase of ca. 30,300 full-resolution images of vouchered leaf specimens vetted to family level, primarily ofangiosperms, including ca. 26,200 cleared and x-rayed leaves from 358 families and ca. 4,100 fossil leaves from49 families. The images are full resolution, with user-friendly filenames, and vetted using APG or modernpaleobotanical standards. The cleared and x-rayed leaves include the Jack A. Wolfe and Leo J. Hickeycontributions to the National Cleared Leaf Collection and a collection of high-resolution scanned x-ray negatives(all of these housed in the Division of Paleobotany, Department of Paleobiology, Smithsonian National Museum ofNatural History), as well as the Daniel Axelrod Cleared Leaf Collection (housed at the University of CaliforniaMuseum of Paleontology, Berkeley). The fossil images include a diverse sampling of Late Cretaceous to Eocenepaleobotanical sites from the Western Hemisphere held at numerous institutions, especially from Florissant FossilBeds National Monument (late Eocene, Colorado), as well as several localities from southern Argentina(Paleocene and Eocene), Colombia (Paleocene), and elsewhere from the Late Cretaceous to Eocene of theWestern USA. The dataset will facilitate new research and education opportunities in paleobotany, comparativeleaf architecture and systematics, and machine learning.