Applications of computer vision and machine learning techniques for digitized herbarium specimens: A systematic literature review

Malik, Owais Ahmed; Ong, Wee-Hong; Slik, Johan Willem Frederik

doi:10.1016/j.ecoinf.2022.101641

Cited by 21 publications

(12 citation statements)

References 82 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our team of seven labelers has logged more than 2000 hours to generate the 494,766 annotations used to train LeafMachine2. To our knowledge, this is the most comprehensive manually annotated training data set for herbarium specimen analysis to date (Hussein et al, 2022 ). We labeled images using an academic license for the Labelbox platform ( https://labelbox.com ), which enabled our labeling team to annotate images remotely, programmatically manage large data sets with the Labelbox API, employ machine‐assisted labeling (MAL), and review labels.…”

Section: Methodsmentioning

confidence: 99%

“…To overcome these limitations, many groups turned to machine learning algorithms, typically some kind of convolutional neural network (CNN), which can categorize individual pixels as members of discrete classes (Ott et al, 2020; Weaver et al, 2020; Younis et al, 2020; Triki et al, 2020, 2021; Goëau et al, 2020, 2022; Guo et al, 2021; Love et al, 2021; Hussein et al, 2021b; Gu et al, 2022; Ott and Lautenschlager, 2022; Milleville et al, 2023). For the task of isolating and measuring individual leaves, semantic segmentation algorithms still lack the power to resolve complex situations (e.g., overlapping leaves) because they produce one mask that contains all leaf pixels and require postprocessing to obtain usable results (Weaver et al, 2020; Hussein et al, 2021b, 2022). Instance segmentation algorithms improve on this as they can directly isolate a single leaf from nearby leaves (Guo et al, 2021; Triki et al, 2021).…”

Section: Term Definitionmentioning

confidence: 99%

“…LeafMachine2 is one of several recently developed tools that are aimed at extracting quantitative trait data from herbarium images (for a summary of the current state of machine learning in a herbarium setting see Hussein et al, 2022 ). Most methods have been semi‐autonomous, requiring some form of human intervention to measure traits.…”

mentioning

confidence: 99%

See 2 more Smart Citations

From leaves to labels: Building modular machine learning networks for rapid herbarium specimen analysis with LeafMachine2

Weaver,

Smith

2023

Appl Plant Sci

View full text Add to dashboard Cite

PremiseQuantitative plant traits play a crucial role in biological research. However, traditional methods for measuring plant morphology are time consuming and have limited scalability. We present LeafMachine2, a suite of modular machine learning and computer vision tools that can automatically extract a base set of leaf traits from digital plant data sets.MethodsLeafMachine2 was trained on 494,766 manually prepared annotations from 5648 herbarium images obtained from 288 institutions and representing 2663 species; it employs a set of plant component detection and segmentation algorithms to isolate individual leaves, petioles, fruits, flowers, wood samples, buds, and roots. Our landmarking network automatically identifies and measures nine pseudo‐landmarks that occur on most broadleaf taxa. Text labels and barcodes are automatically identified by an archival component detector and are prepared for optical character recognition methods or natural language processing algorithms.ResultsLeafMachine2 can extract trait data from at least 245 angiosperm families and calculate pixel‐to‐metric conversion factors for 26 commonly used ruler types.DiscussionLeafMachine2 is a highly efficient tool for generating large quantities of plant trait data, even from occluded or overlapping leaves, field images, and non‐archival data sets. Our project, along with similar initiatives, has made significant progress in removing the bottleneck in plant trait data acquisition from herbarium specimens and shifted the focus toward the crucial task of data revision and quality control.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Term Definitionmentioning

confidence: 99%

See 1 more Smart Citation

From leaves to labels: Building modular machine learning networks for rapid herbarium specimen analysis with LeafMachine2

Weaver,

Smith

2023

Appl Plant Sci

View full text Add to dashboard Cite

show abstract

“…Machine learning and artificial intelligence has been successfully applied to digital herbarium specimens in applications such as species identification (Soltis et al 2020), the identification of rare and red listed taxa (Rocchetti et al 2021), the identification of phenology stages of the specimens (Pearson et al 2020), to extract leaf trait data (Weaver et al 2020), and to identify types and levels of herbivory (Meineke et al 2020). Such technological advances can be applied to digital Arctic specimens and can vastly advance our ability to inform a multitude of evolutionary and ecological studies in a rapidly changing landscape (Hussein et al 2022).…”

Section: Arctic Specimens For Global Change Research and Nature Manag...mentioning

confidence: 99%

Digitization of the Greenland Vascular Plant Herbarium as a Unique Research Infrastructure to Study Arctic Climate Change and Inform Nature Management

Ahlstrand

2023

Collections: A Journal for Museum and Archives Professionals

View full text Add to dashboard Cite

The Greenland Vascular Plant Herbarium at the Natural History Museum of Denmark represents the largest collection of botanical specimens from Greenland and includes some of the oldest known specimens collected in the Arctic, as well as voucher specimens collected over time from important botanical expeditions. High-resolution digital images for all specimens in this collection have recently been obtained and accompanying specimen label data have been transcribed. Digitizing this invaluable botanical collection from Greenland allows us to make nearly 170,000 Arctic plant specimens available online to researchers, amateur botanists, nature managers and advisers, as well as the general public. Improved access to this museum collection will facilitate global change research and nature management in Greenland’s rapidly changing Arctic environment and will help promote the value of digitizing Arctic specimens maintained in natural history collections worldwide. Current and potential applications of Arctic herbarium material in climate change studies, and the biases and limitations of such herbarium material for these studies are discussed.

show abstract

“…Finally, machine learning algorithms are starting to be developed for automated trait assessments, as well as for improved modelling of population genetic patterns (e.g. Hussein et al, 2021; Schrider & Kern, 2018).…”

Section: Challenges Associated With Historical Collections and How To...mentioning

confidence: 99%

Integration of historic collections can shed light on patterns of change in plant–pollinator interactions and pollination service

et al. 2022

View full text Add to dashboard Cite

Mutualistic interactions between plants and animal pollinators are increasingly under threat through anthropogenic change, and it is critical to understand how temporal changes affect the structure and function of these ecologically important interactions. Because the responses of plant–pollinator interactions to anthropogenic change may take place over decades, historical collections that store information across long time horizons contribute uniquely to our understanding. In this article, we highlight several key questions related to long‐term changes in the structure and function of plant–pollinator interactions. We articulate how research could proceed rapidly via new techniques, greater integration of resources in museum collections along with coincident use of a single data source. We acknowledge the challenges that come with using historical collections and discuss how to minimize them. We provide suggestions that will allow for full utilization of museum resources for addressing a variety of issues regarding plant–pollinator interactions. This perspective paper aims to stimulate new integrative research aimed at understanding temporal patterns in plant–pollinator interactions. Read the free Plain Language Summary for this article on the Journal blog.

show abstract

Applications of computer vision and machine learning techniques for digitized herbarium specimens: A systematic literature review

Cited by 21 publications

References 82 publications

From leaves to labels: Building modular machine learning networks for rapid herbarium specimen analysis with LeafMachine2

From leaves to labels: Building modular machine learning networks for rapid herbarium specimen analysis with LeafMachine2

Digitization of the Greenland Vascular Plant Herbarium as a Unique Research Infrastructure to Study Arctic Climate Change and Inform Nature Management

Integration of historic collections can shed light on patterns of change in plant–pollinator interactions and pollination service

Contact Info

Product

Resources

About