Littérature scientifique sur le sujet « Herbarium scans »

As herbarium specimens are increasingly becoming digitised and accessible in online repositories, advanced computer vision techniques are being used to extract information from them. The presence of certain plant organs on herbarium sheets is useful information in various scientific contexts and automatic recognition of these organs will help mobilise such information. In our study, we use deep learning to detect plant organs on digitised herbarium specimens with Faster R-CNN. For our experiment, we manually annotated hundreds of herbarium scans with thousands of bounding boxes for six types of plant organs and used them for training and evaluating the plant organ detection model. The model worked particularly well on leaves and stems, while flowers were also present in large numbers in the sheets, but were not equally well recognised.

A.K. Skvortsov is renowned among botanists both in this country and abroad as a high-ranking specialist in plant systematics, the study of floras, and plant introduction and acclimation. Scientific interests and professional activities of A.K. Skvortsov included the flora of Eastern Europe; systematics of the genera Salix L., Betula L., Populus L., and Epilobium Dill. ex L.; general problems of the evolutionary theory, intraspecific variability, and microevolution. A.K. Skvortsov made a large contribution to the development of herbarium collections. Within the span of 65 years (1938–2002) he was collecting during expeditions and field trips across Russia, Europe, North America, India, and China, his collections amounting to more than 50 thousand sheets. From 1966, Skvortsov curated the herbarium collection at the Main Botanic Garden RAN (MHA). Here his personal herbarium collection is preserved along with his specimens from temperate European Russia and Lower Volga and separate collections of willows, poplars, and birches. The book by Skvortsov Herbarium methods and techniques. A manual (1977) has served a handbook for a few generations of botanists. A.K. Skvortsov defined the role of herbaria in the contemporary science as follows: “Full and reliable data on dynamics of the flora in any concrete country during a certain time period can be provided only by the herbarium collection. Versatile application potential, multi-functionality is an essential, immensely important quality of a herbarium specimen. With the advancement of science, it is possible to extract more and more information from the same herbarium specimen. Hence a herbarium specimen is a primary, authentic document that cannot be substituted by any secondary, derived kind of documentation (Skvortsov 1977: 4). Specimens exchange is an old tradition of botanical institutions. In his description of the Herbarium of the Main Botanical Garden in Moscow (2005), Skvortsov listed the major external sources of its enhancement, which included the Central Siberian Botanical Garden, Siberian Branch RAN (CSBG SB RAN) in Novosibirsk, from where MHA had received more than 13 thousand specimens collected in different parts of Siberia. At the same time, the Central Siberian Botanical Garden was receiving material from the Moscow Main Botanical Garden in exchange. The herbarium collections of the CSBG SB RAN (NS and NSK) currently amount to about 680 thousand specimens of vascular plants. Of these, 37 500 have been scanned with ObjectScan 1600 (Microtek) scanners, in accordance with the international standards: optical resolution of 600 dpi, scans accompanied with the color guide and linear scale ruler, barcoded, and preserved within the online Virtual Herbarium at the open-access site of the CSBG SB RAN (http://herb.csbg.nsc.ru:8081). Information extracted from herbarium labels and scans of herbarium specimens are also published as seven datasets on the GBIF (Global Botanical Information Facility) Portal (gbif.org). Among the scanned specimens, there are collections by A.K. Skvortsov from Volgograd, Belgorod, Bryansk, Smolensk, Oryol, Moscow, Sverdlovsk, and Amur regions as well as Crimea and Karachay-Cherkess Republic, Estonia, Armenia, Georgia, Ukraine, Kazakhstan, Sweden, India – a total of 450 sheets received through the exchange between the MHA and NS/NSK. A.K. Skvortsov made an invaluable contribution to the herbarium business development.  

A better knowledge of tree vegetative growth phenology and its relationship to environmental variables is crucial to understanding forest growth dynamics and how climate change may affect it. Less studied than reproductive structures, vegetative growth phenology focuses primarily on the analysis of growing shoots, from buds to leaf fall. In temperate regions, low winter temperatures impose a cessation of vegetative growth shoots and lead to a well-known annual growth cycle pattern for most species. The humid tropics, on the other hand, have less seasonality and contain many more tree species, leading to a diversity of patterns that is still poorly known and understood. The work in this study aims to advance knowledge in this area, focusing specifically on herbarium scans, as herbariums offer the promise of tracking phenology over long periods of time. However, such a study requires a large number of shoots to be able to draw statistically relevant conclusions. We propose to investigate the extent to which the use of deep learning can help detect and type-classify these relatively rare vegetative structures in herbarium collections. Our results demonstrate the relevance of using herbarium data in vegetative phenology research as well as the potential of deep learning approaches for growing shoot detection.

Global digitization efforts have archived millions of specimen scans worldwide in herbarium collections, which are essential for studying plant evolution and biodiversity. ReColNat hosts, at present, over 10 million images. However, analyzing these datasets poses crucial challenges for botanical research. The application of deep learning in biodiversity analyses, particularly in analyzing herbarium scans, has shown promising results across numerous tasks (Ariouat et al. 2023, Ariouat et al. 2024, Groom et al. 2023, Sahraoui et al. 2023).Within the e-Col+project (ANR-21-ESRE-0053), we are developing multiple deep learning models aimed at identifying plant morphological traits. We have developed pipelines and models for cleaning, analyzing, and transforming herbarium images, including models for: i) detecting non-vegetal elements, such as barcodes, envelopes, labels, etc.; ii) detecting plant organs, including leaves, flowers, fruits, etc.; and iii) segmenting to recognize plant parts for image cleaning. We are also developing models for classification tasks related to various morphological traits.To validate these models, improve their generalization, and make them easily usable by end-users, deploying them within a generic platform is crucial. The generic platform called PlantAI, currently under development by the e-Col+ project, should enable easy deployment during development for testing and allow users to load annotations for new traits in order to train a model and add it to the existing catalog. The platform is based on a microservice architecture, allowing users to upload images, create custom datasets, and access various AI models for image analysis.The platform is composed of four main modules, as illustrated in Fig. 1. The first module is the collaborative workspace manager, which allows users to create projects and image datasets and invite other users to collaborate on a project. The second module is the navigation interface and dashboards. This module integrates a search engine using metadata and AI annotations, a navigation interface between projects, datasets, and specimens, as well as dashboards for analysis across datasets, specimens, and AI models.The third module is the dataset manager, which handles metadata and annotations associated with the specimens. These annotations can be produced either by expert users or by AI models. The fourth module is the AI models management module, so that models can be used to generate AI annotations of specimen. During the development lifecycle of an AI model, users can create datasets and annotate them with AI models. These annotations can be in two possible states: validated by experts and non-validated. Users collaborating on a project can indicate errors in the model predictions and leave comments to explain their evaluations. These corrections made by experts can be used to retrain the models and thus improve their performance.This platform, will be highly beneficial for botanists, enhancing the efficiency and effectiveness of biodiversity analyses from herbarium scans. We aim to provide users with a catalog of AI models through this platform and allow them to import their own datasets with their own annotations regarding traits of their choice. Users will be able to select a model from the AI model catalog and train it using their dataset. Ultimately, the model obtained from this training will be automatically deployed to be available for AI annotation. The annotations produced by this model will be automatically available in the filtering and navigation interface, thus allowing for dynamic and automatic integration of the AI annotations into the navigation interface.

High-resolution scans of plant cuttings were made for a plant identification course to create additional study resources. Stems, flowers, leaves, and other parts with identifiable features were cut and placed on a high-quality flatbed scanner. A framework suspended a black background cloth above the cuttings to create a dark scanning environment, and it was placed far enough away from the scanner glass so as not to appear in the scanned image. Botanical scans can be shared, manipulated, composed, and otherwise provided to students for study materials. Scans are complementary to other common study aids such as pressed herbarium samples or photography.

Herbarium specimen scans constitute a valuable source of raw data. Herbarium collections are gaining interest in the scientific community as their exploration can lead to understanding serious threats to biodiversity. Data derived from scanned specimen images can be analyzed to answer important questions such as how plants respond to climate change, how different species respond to biotic and abiotic influences, or what role a species plays within an ecosystem. However, exploiting such large collections is challenging and requires automatic processing. A promising solution lies in the use of computer-based processing techniques, such as Deep Learning (DL). But herbarium specimens can be difficult to process and analyze as they contain several kinds of visual noise, including information labels, scale bars, color palettes, envelopes containing seeds or other organs, collection-specific barcodes, stamps, and other notes that are placed on the mounting sheet. Moreover, the paper on which the specimens are mounted can degrade over time for multiple reasons, and often the paper's color darkens and, in some cases, approaches the color of the plants.Neural network models are well-suited to the analysis of herbarium specimens, while making abstraction of the presence of such visual noise. However, in some cases the model can focus on these elements, which eventually can lead to a bad generalization when analyzing new data on which these visual elements are not present (White et al. 2020). It is important to remove the noise from specimen scans before using them in model training and testing to improve its performance. Studies have used basic cropping techniques (Younis et al. 2018), but they do not guarantee that the visual noise is removed from the cropped image. For instance, the labels are frequently put at random positions into the scans, resulting in cropped images that still contain noise. White et al. (2020) used the Otsu binarization method followed by a manual post-processing and a blurring step to adjust the pixels that should have been assigned to black during segmentation. Hussein et al. (2020) used an image labeler application, followed by a median filtering method to reduce the noise. However, both White et al. (2020) and Hussein et al. (2020) consider only two organs: stems and leaves. Triki et al. (2022) used a polygon-based deep learning object detection algorithm. But in addition to being laborious and difficult, this approach does not give good results when it comes to fully identifying specimens. In this work, we aim to create clean high-resolution mask extractions with the same resolution as the original images. These masks can be used by other models for a variety of purposes, for instance to distinguish the different plant organs. Here, we proceed by combining object detection and image segmentation techniques, using a dataset of scanned herbarium specimens. We propose an algorithm that identifies and retains the pixels belonging to the plant specimen, and removes the other pixels that are part of non-plant elements considered as noise. A removed pixel is set to zero (black). Fig. 1 illustrates the complete masking pipeline in two main stages, object detection and image segmentation.In the first stage, we manually annotated the images using bounding boxes in a dataset of 950 images. We identified (Fig. 2) the visual elements considered to be noise (e.g., scale-bar, barcode, stamp, text box, color pallet, envelope). Then we trained the model to automatically remove the noise elements. We divided the dataset into 80% training, 10% validation and 10% test set. We ultimately achieved a precision score of 98.2%, which is a 3% improvement from the baseline. Next, the results of this stage were used as input for image segmentation, which aimed to generate the final mask. We blacken the pixels covered by the detected noise elements, then we used HSV (Hue Saturation Value) color segmentation to select only the pixels with values in a range that corresponds mostly to a plant color. Finally, we applied the morphological opening operation that removes noise and separates objects; and the closing operation that fills gaps, as described in Sunil Bhutada et al. (2022) to remove the remaining noise. The output here is a generated mask that retains only the pixels that belong to the plant. Unlike other proposed approaches, which focus essentially on leaves and stems, our approach covers all the plant organs (Fig. 3). Our approach removes the background noise from herbarium scans and extracts clean plant images. It is an important step before using these images in different deep learning models. However, the quality of the extractions varies depending on the quality of the scans, the condition of the specimens, and the paper used. For example, extractions made from samples where the color of the plant is different from the color of the background were more accurate than extractions made from samples where the color of the plant and background are close. To overcome this limitation, we aim to use some of the obtained extractions to create a training dataset, followed by the development and the training of a generative deep learning model to generate masks that delimit plants.

Hundreds of herbarium collections have accumulated a valuable heritage and knowledge of plants over several centuries (Page et al. 2015). Recent initiatives, such as iDigBio (https://www.idigbio.org), aggregate data from and images of vouchered herbarium sheets (and other biocollections) and make this information available to botanists and the general public worldwide through web portals. These ambitious plans to transform and preserve these historical biodiversity data into digital format are supported by the United States National Science Foundation (NSF) Advancing the Digitization of Natural History Collections (ADBC) and the digitization is done by the Thematic Collections Networks (TCNs) funded under the ADBC program. However, thousands of herbarium sheets are still unidentified at the species level while numerous sheets should be reviewed and updated following more recent taxonomic knowledge. These annotations and revisions require an unrealistic amount of work for botanists to carry out in a reasonable time (Bebber et al. 2010). Computer vision and machine learning approaches applied to herbarium sheets are promising (Wijesingha and Marikar 2012) but are still not well studied compared to automated species identification from leaf scans or pictures of plants taken in the field. In a recent study, we evaluate the accuracy with which herbarium images can be potentially exploited for species identification with deep learning technology (Carranza-Rojas et al. 2017), particularly Convolutional Neural Networks (CNN) (Szegedy et al. 2015). This type of network allows automatic learning of the most prominent visual patterns in the images since they are trainable end-to-end (thus, differentiable), as opposed to previous approaches that use custom, hand-made feature extractors. A first challenge is to use herbarium sheet images alone to automatically identify the species of plants mounted on herbarium sheets. Secondly, we propose studying if the combination of herbarium sheet images with photos of plants in the field (Joly et al. 2015, Carranza-Rojas and Mata-Montero 2016) is a viable idea to train models that provide accurate results during identification. Finally, we explore if herbarium images from one region with a specific flora can be used in transfer learning (a technique in deep learning that first allows training a model with a dataset and then once trained, uses the weighted results to train another model with that knowledge as the baseline) to another region with other species; for example, in a region under-represented in terms of collected data. Our evaluation shows that the accuracy for species identification with deep learning technology, based on herbarium images, reaches 90.3% on a dataset of more than 1200 European plant species. This could potentially lead to the creation of a semi-, or even fully automated system to help taxonomists and experts with their annotation, classification, and revision works. In this paper, we take a closer look at the accuracy levels achieved with respect to the first two challenges. We evaluate the accuracy levels for each species included in the dataset, which encompasses 253,733 images, 1,204 species.

The Herbarium at the Royal Botanic Garden Edinburgh (RBGE) have run Frankenstein's Plants as part of the Edinburgh Science Festival, which runs during the Easter holidays. Frankenstein's Plants aims to engage both children and the wider public with the work of herbaria, highlighting the Herbarium at RBGE. We first took part in 2019, and continued in 2023 and 2024.The Royal Botanic Garden Edinburgh is a well-known and loved location for visitors and local residents, attracting large numbers of people throughout the year. However, the work of the Herbarium is less well-known and many visitors, including residents, are unaware of the Herbarium collection.The aims of the event include:Raising awareness of the herbarium at RGBE, and how it is used by scientists globallyEducating people on how specimens are madeIntroducing the idea of a scientific name and how it is constructedFrankenstein's Plants takes participants through the 'life' of a specimen, starting with the selection of material for mounting, through to the digitisation of the specimen, contributing to a virtual herbarium. Participants are encouraged to let their imagination go wild and create a monster (or whatever else they feel like) using the material provided. The event is laid out with a series of stations where staff talk through and support each step of the process (Fig. 1).The key steps of specimen creation for this event are:Select plant material: pressed and dried prior to the event, using flowers and foliage bought from an online florist, alongside material gathered from the gardens at RBGE.Mount the specimen: a pre-printed label is attached to a piece of board. The boards we use are approximately A4, allowing enough room for participants to create their creatures. Gummed tape is used to fix the plant material to the sheets, as a relatively low-mess option. The label provides space for recording the species name, a description and 'collector' information. Some locality information and a barcode is prefilled. The details on the label aim to give the participant an idea of the types of data that would typically be recorded when collecting.Name the specimen: participants can create their own name for their specimen. A list of options for the genus and species is provided and consists of both real and made-up genera and species epithets. This is an opportunity for the team to talk about how plants are named using the binomial system, and the importance of Latin names for communicating about life on earth.Describe the specimen: basic botanical terms are provided alongside sketches to get the participants thinking about how species can be described. This step can be modified based on the age of the participant, bringing in more information and technical terms where appropriate.Stamp the specimen: whilst not a key part of the processing of a specimen, it is very much enjoyed by the participants! We have a mix of old stamps that were previously used on specimens that they can choose from.Digitise the specimen: the final step is to take a picture of the specimen. We have a copy stand and DSLR (digital single-lens reflex) camera to allow the participant to take a photo of their specimen. The participant scans the barcode, which we use for the file name and later to look up the newly-digitised specimen in the virtual herbarium. They then take the photo, using remote shutter release triggered by a fixed mouse. The laptop is connected to a big screen, so everyone can see the image of their specimen.Following the event all the images are uploaded to a Flickr gallery, which acts as our virtual herbarium. A link to this gallery is provided on the label, allowing participants to look up their specimen online, as well as providing an approximate count of participants at the event.We have members of staff and volunteers from the PhD and MSc cohorts at the RBGE to talk through each step, as well as providing information about the work of the herbarium and Science Division in general. This is a great opportunity for us to engage with the public about our work, especially as the herbarium team has limited regular engagement with the wider public.Alongside the event we have a display about the Herbarium, including teaching specimens. The display provides information about the collections, without needing a team member to explain it, although we aim to have a member of staff available. Often, we find it is the parents/guardians of the children who are most interested in this, so it allows us to extend the group of people with whom we are engaging.Over the three years we have run the event, we have engaged with over 1,000 'collectors', ranging in age from 3-95 (Fig. 2), comprised of people from Edinburgh and visitors to the garden from farther afield. This figure does not include those additional members of the public who attend the event, as e.g. the parent/guardian of a child, but do not engage in creating a specimen themselves. If we include an estimate of these additional members of the public, the number of people engaged is nearer to 2,000–3,000.Feedback, both during the event and post-event, has been exceptional. We have had comments on how engaging the event is, as well as requests to take pictures so that others can run a similar event themselves.

A taxonomic revision is presented of the Ecuadorian species of Sciodaphyllum P. Browne (Araliaceae) belonging to the Violaceum group, characterized by leaves that often dry with a distinctive purplish or brownish-purple cast, the presence of a long, clawlike, generally adaxially incurved stipular ligule, and flowers borne in umbellules. Analysis of herbarium material supplemented by field observations has revealed a total of 11 species, all but one of which are described as new. This brings the total number of published Sciodaphyllum occurring in Ecuador to 47. An identification key is provided; each species is mapped and illustrated by scans of the type collection, accompanied by color photos taken in the field of four of the new taxa. Preliminary extinction risk assessments conducted using the IUCN Red List criteria indicate that one species is Critically Endangered (CR), four are Endangered (EN), and three are Vulnerable (VU), while three are regarded as Least Concern (LC).

International mass digitization efforts through infrastructures like the European Distributed System of Scientific Collections (DiSSCo), the US resource for Digitization of Biodiversity Collections (iDigBio), the National Specimen Information Infrastructure (NSII) of China, and Australia's digitization of National Research Collections (NRCA Digital) make geo- and biodiversity specimen data freely, fully and directly accessible. Complementary, overarching infrastructure initiatives like the European Open Science Cloud (EOSC) were established to enable mutual integration, interoperability and reusability of multidisciplinary data streams including biodiversity, Earth system and life sciences (De Smedt et al. 2020). Natural Science Collections (NSC) are of particular importance for such multidisciplinary and internationally linked infrastructures, since they provide hard scientific evidence by allowing direct traceability of derived data (e.g., images, sequences, measurements) to physical specimens and material samples in NSC. To open up the large amounts of trait and habitat data and to link these data to digital resources like sequence databases (e.g., ENA), taxonomic infrastructures (e.g., GBIF) or environmental repositories (e.g., PANGAEA), proper annotation of specimen data with rich (meta)data early in the digitization process is required, next to bridging technologies to facilitate the reuse of these data. This was addressed in recent studies (Younis et al. 2018, Younis et al. 2020), where we employed computational image processing and artificial intelligence technologies (Deep Learning) for the classification and extraction of features like organs and morphological traits from digitized collection data (with a focus on herbarium sheets).However, such applications of artificial intelligence are rarely—this applies both for (sub-symbolic) machine learning and (symbolic) ontology-based annotations—integrated in the workflows of NSC's management systems, which are the essential repositories for the aforementioned integration of data streams.This was the motivation for the development of a Deep Learning-based trait extraction and coherent Digital Specimen (DS) annotation service providing "Machine learning as a Service" (MLaaS) with a special focus on interoperability with the core services of DiSSCo, notably the DS Repository (nsidr.org) and the Specimen Data Refinery (Walton et al. 2020), as well as reusability within the data fabric of EOSC. Taking up the use case to detect and classify regions of interest (ROI) on herbarium scans, we demonstrate a MLaaS prototype for DiSSCo involving the digital object framework, Cordra, for the management of DS as well as instant annotation of digital objects with extracted trait features (and ROIs) based on the DS specification openDS (Islam et al. 2020). Source code available at: https://github.com/jgrieb/plant-detection-service