Applications of deep convolutional neural networks to digitized natural history collections

By

Schuettpelz, Eric, 1977-
Frandsen, Paul
Dikow, Rebecca
Brown, Abel
Orli, Sylvia
Peters, Melinda
Metallo, Adam
Funk, V. A. (Vicki Ann), 1947-
Dorr, Laurence

Type

Article

Date of Publication

2017-11-02

Original Publication

Biodiversity data journal

Volume

5

Pages

1--9

Subjects

convolutional neural networks , deep learning , machine learning , Mass Digitization , Natural history collections

Contributed by

Pensoft Publishers

Language

English

Published by

Sofia Pensoft Publishers [2013]-

Identifiers

DOI: https://doi.org/10.3897/BDJ.5.e21139
eLocator: e21139

 

Abstract

Natural history collections contain data that are critical for many scientific endeavors. Recent efforts in mass digitization are generating large datasets from these collections that can provide unprecedented insight. Here, we present examples of how deep convolutional neural networks can be applied in analyses of imaged herbarium specimens. We first demonstrate that a convolutional neural network can detect mercury-stained specimens across a collection with 90% accuracy. We then show that such a network can correctly distinguish two morphologically similar plant families 96% of the time. Discarding the most challenging specimen images increases accuracy to 94% and 99%, respectively. These results highlight the importance of mass digitization and deep learning approaches and reveal how they can together deliver powerful new investigative tools.