Taxon and trait recognition from digitized herbarium specimens using deep convolutional neural networks

TL;DR

This paper demonstrates that deep convolutional neural networks can effectively identify plant taxa and traits from digitized herbarium specimen images, promising advancements in taxonomy and collection management.

Contribution

It introduces a CNN-based method for recognizing taxa and traits from herbarium images, combining image data with morphological traits for improved accuracy.

Findings

High accuracy in species recognition

Effective trait identification from images

Potential for integration into taxonomic tools

Abstract

Herbaria worldwide are housing a treasure of 100s of millions of herbarium specimens, which are increasingly being digitized in recent years and thereby made more easily accessible to the scientific community. At the same time, deep learning algorithms are rapidly improving pattern recognition from images and these techniques are more and more being applied to biological objects. We are using digital images of herbarium specimens in order to identify taxa and traits of these collection objects by applying convolutional neural networks (CNN). Images of the 1000 species most frequently documented by herbarium specimens on GBIF have been downloaded and combined with morphological trait data, preprocessed and divided into training and test datasets for species and trait recognition. Good performance in both domains is promising to use this approach in future tools supporting taxonomy and natural history collection management.