Generating Masks for Image Segmentation in Digitized Herbarium Specimens

Digitized herbarium images contain complex information unrelated to the shape and color of the specimens represented within them. This information can contribute a substantial amount of noise if one is to use the image as a proxy for pattern, shape, or color of the specimen. Image segmentation, whereby the specimen material is partitioned from the background (e.g., herbarium sheet, label, color ramp), offers one possible solution, yet training data for image segmentation of herbarium specimens is nonexistent. We present a pipeline for generating training data for image segmentation tasks along with a novel dataset of highly resolved image masks segmenting plant material from background noise. This dataset can be used to train neural networks to segment plant material in herbarium sheets more generally, and our method is applicable to other museum data sources where masking may be useful for quantitative analysis of patterns and shapes

Use of Semantic Segmentation for Increasing the Throughput of Digitisation Workflows for Natural History Collections

The need to increase global accessibility to specimens while preserving the physical specimens by reducing their handling motivates digitisation. Digitisation of natural history collections has evolved from recording of specimens’ catalogue data to including digital images and 3D models of specimens. The sheer size of the collections requires developing high throughput digitisation workflows, as well as novel acquisition systems, image standardisation, curation, preservation, and publishing. For instance, herbarium sheet digitisation workflows (and fast digitisation stations) can digitise up to 6,000 specimens per day; operating digitisation stations in parallel can increase that capacity. However, other activities of digitisation workflows still rely on manual processes which throttle the speed with which images can be published. Image quality control and information extraction from images can benefit from greater automation. This presentation explores the advantages of applying semantic segmentation (Fig. 1) to improve and automate image quality management (IQM) and information extraction from images (IEFI) of physical specimens. Two experiments were designed to determine if IQM and IEFI activities can be improved by using segments instead of full images. The time for segmenting full images needs to be considered for both IQM and IEFI. A semantic segmentation method developed by the Natural History Museum (Durrant and Livermore 2018) adapted for segmenting herbarium sheet images (Dillen et al. 2019) can process 50 images in 12 minutes. The IQM experiments evaluated the application of three quality attributes to full images and to image segments: colourfulness (Fig. 2), contrast (Fig. 3) and sharpness (Fig. 4). Evaluating colourfulness is an alternative to colour quantization algorithms such as RMSE and Delta E (Hasler and Suesstrunk 2003, Palus 2006), the method produces a value indicating if the image degrades after processing. Contrast measures the difference in luminance or colour that makes an object distinguishable. Contrast is determined by the difference in colour and brightness of the object and other objects within the same field of view (Matkovic et al. 2005, Präkel 2010). Sharpness encompasses the concepts of resolution and acutance (Bahrami and Kot 2014, Präkel 2010). Sharpness influences specimen appearance and readability of information from labels and barcodes. Evaluating the criteria on 56 barcodes and 50 colour charts segments extracted from fifty images took 34 minutes (8 minutes for the barcodes and 26 minutes for colour charts). The evaluation on the corresponding full images took 100 minutes. The processing of individual segments and full images provided results equivalent to subjective manual quality management. The IEFI experiments compared the performance of four optical character recognition (OCR) programs applied to full images (Drinkwater et al. 2014) against individual segments. The four OCR programs evaluated were Tesseract 4.X, Tesseract 3.X, Abby FineReader Engine 12, and Microsoft OneNote 2013. The test was based on a set of 250 herbarium sheet images and 1,837 segments extracted from them. The results from the experiments show that there is an average OCR speed-up of 49% when using segmented images when compared to processing times for full images (Table 1). Similarly, there was an average increase of 13% in line correctness (information from lines is ordered and not fragmented (Fig. 5, Table 2 ). Additionally, the results are useful for comparing the four OCR programs, with Tesseract 3.x offering shortest processing time, while Tesseract 4.X achieving the highest scores for line accuracy (including hand written text recognition). The results suggest that IEFI could be improved by performing OCR using segments rather than whole images, leading to faster processing and more accurate outputs. The findings support the feasibility of further automation of digitisation workflows for natural history collections. In addition to increasing the accuracy and speed of IQM and IEFI activities, the explored approaches can be packaged and published, enabling automated quality management and information extraction to be offered as a service, taking advantage of cloud platforms and workflow engines.

Fagonia schweinfurthii (Hadidi) Hadidi ex Ghafoor (Zygophyllaceae) – An Addition to the flora of Uttar pradesh, India

Fagonia schweinfurthii (Zygophyllaceae) is reported here as an addition to the flora of Uttar Pradesh, collected from Chambal Wildlife Sanctuary, Etawah. A detailed description and scanned herbarium sheet image are provided to facilitate proper identification.

Engaging Children with Herbarium Specimens at Royal Botanic Gardens Kew’s Science Festival

The Kew and Wakehurst Science Festivals consists of five days of activities over two weekends. Workshops and tours allow visitors to engage with the scientists and their research. We designed an interactive experience, so children could understand what a herbarium sheet is and the process of making one. The Herbarium accessions an average of 30,000 specimens per year and because specimens need to have a long life and be able to withstand being handled for hundreds of years they need to be ‘mounted’ according to strict protocols and guidelines. Botanical specimens are vital to research at Kew and beyond, providing key scientific data. Once mounted onto herbarium sheets botanical specimens are added to the Herbarium and made widely available to visiting scientists and researchers. Digitising these specimens increases access further through online portals. To achieve a specimen that can be handled for many years the specimens are mounted onto archival paper, along with their labels, before being added to the collection. There are 6 members in RBG Kew’s Specimen Preparation team who work full time to prepare botanical specimens for accession into the Herbarium collection; which currently stands at 7 million specimens and the oldest dates from the 1700s. We simplified this specimen preparation process down to the basic component parts of paper, glue, plant material and pressing. Using material and tools that visitors would be able to find for themselves; art paper, child friendly glue and plant material used in flower crafts we created a hands-on experience for mounting a herbarium specimen. The Science Festival is now in its 3rd year and each year the activity has been modified based on lessons learned over the course of the festival and each year. The stall is immensely popular going from 300 participants in the first year to over 700 in 2017. In the second year we added a new dimension and allowed visitors to image the specimens they created allowing them to zoom in and see plant parts and structures in further detail to highlight the importance of digitisation. These images can be viewed on the Kew Science Flickr group.

Putting your Finger upon the Simplest Data

Over the past decades, digitization endeavors across many institutions holding natural history collections (NHCs) have multiplied with three broad aims: first, to facilitate collection management by moving existing analog catalogues into digital form; second, to efficiently document and inventory specimens in collections, including imaging them as taxonomical surrogates; and third, to enable discovery of, and access to, the resulting collection data. NHCs contain a unique wealth of potential knowledge in the form of primary biodiversity data records (PBR): at its most basic level, the “what, where and when” of occurrences of the specimens in the collections. But as T.S. Eliot famously said, “knowledge is invariably a matter of degree”. For such data to be transformed into digitally accessible knowledge (DAK) that is conducive to an understanding about how the natural world works, release of digitized data (the “this we know”) is necessary. At least two billion specimens are estimated to exist in NHCs already, but only a small fraction can be considered properly DAK: most have either not been digitized yet, or not released through a discovery facility. Digitizing is relatively costly as it often entails manually processing each specimen unit (e.g. a herbarium sheet, a pinned insect, or a vial full of invertebrates). How long could it take us to transform all NHCs into DAK? Can we keep up with the natural growth in collections? The Global Biodiversity Information Facility (GBIF) has become the de facto main index of PBR, both originated in NHCs or as field observations. Digitized NHC that are standards-compliant and can be connected to, or harvested by, GBIF, effectively become DAK. I have examined GBIF growth data looking for a pattern of DAK generation. I found that the rate of NHC-based PBR accrual is remarkably constant: the total DAK shows a strongly linear growth, as opposed to the exponential growth exhibited by cumulative observation data. Projecting the trend to the estimated holdings shoots the completion many decades ahead. In addition, digitized data appear to be taxonomically biased. Digitization efforts must therefore step up qualitatively in order to enable processing the backlog, let alone newly-acquired accessions, within one generation. Among several possible solutions, emerging, industrial-scale mass-digitization techniques may help harnessing this otherwise daunting task—but there’s also a risk that DAK becomes even more uneven across taxon groups because of the narrow application specificity of such techniques, thus potentially biasing our knowledge of nature.

Typification of the name Botrychium lanuginosum (Ophioglossaceae)

The genus Botrychium Swartz (1801: 110) is represented by 50–60 species, distributed worldwide (Zhang et al. 2013). In India, the genus is represented by 8 taxa (Fraser- Jenkins et al. 2017). During revisionary study of the genus Botrychium in India, the authors realized the necessity to lectotypify the name Botrychium lanuginosum Wallich ex Hooker & Greville (1828: t.79) because no specific herbarium sheet is cited as holotype in the protologue and has not been lectotypified earlier (Fraser-Jenkins et al. 2017). Therefore it is lectotypified here to avoid any ambiguity in the application of the name. The guidelines of Art. 9.2 and Recommendations 9A, 9C and 9D of the ICN (McNeill et al. 2012) have been followed while designating the lectotype

Typification of thirty six names of thirty five recognized taxa in Impatiens (Balsaminaceae), endemic to Western Ghats

The genus Impatiens Linnaeus (1753: 937) consists of about 1000 species, distributed in the tropical and subtropical regions of the Old World as well as in the northern temperate regions (Mabberley 2008). In India there are about 215 species of Impatiens recorded so far, of which 142 species are endemic, and the diversity occurs in two regions, viz. the Eastern Himalayas and the Western Ghats. During systematic study of this genus in the Western Ghats region of India, I realized that it was necessary to typify 36 binomials of 35 recognized endemic species from the Western Ghats, because no specific herbarium sheet was cited as holotype in protologue of these names and also yet not typified (Vivekananthan et al. 1997, Rathakrishnan 2005, Bhaskar 2012). Here lectotypes are chosen for 35 names and a neotype for one name. The images of the lectotypes, isolectotypes, residual syntypes and neotype specimens at BR, E, G, K, M, MPU, NY and P are available online (http://www.br.fgov.be/research/collections/herbarium, http://elmer.rbge.org.uk/bgbase/vherb/bgbasevherb.php, http://www.ville-ge.ch/musinfo/bd/cjb/chg, http://apps.kew.org/herbcat/navigator.do, https://plants.jstor.org, http://www.collections.univ-montp2.fr/herbier-mpu-presentation/base-de-donnees-botanique-herbier-mpu, http://sweetgum.nybg.org/science/vh & https://science.mnhn.fr/institution/mnhn/search/form), while those at BSI, CAL and MH are not available online. Therefore, the images of selected lectotypes from MH are provided here.

Typification of Dioscorea communis and its synonym Tamus communis var. subtriloba (Dioscoreaceae)

The typification of Dioscorea communis (≡ Tamus communis, Dioscoreaceae) is discussed. The designation of the nomenclatural type is based on the consultation of Linnaeus’s original material preserved at LINN, the Clifford Herbarium at BM, the Herbarium Joachim Burser at UPS, and the literature cited in the protologue. The name is lectotypified using a herbarium sheet preserved at LINN. A lectotype is also selected for the name Tamus communis var. subtriloba, from a specimen preserved at the Gussone’s herbarium at NAP.

Remarks and clarifications on the type material of Ximenesia encelioides (Verbesina encelioides; Asteraceae, Heliantheae), with a new lectotype designation

In a recent paper, the name Ximenesia encelioides was lectotypified on material conserved at MA. However, the selected “lectotype” includes several fragments collected at three different times, and that herbarium sheet is therefore to be considered to contain three different specimens as defined in Art. 8.2 of the International Code of Nomenclature of algae, fungi and plants. This fact makes that “lectotype” invalid since contrary to Art. 8.1, and hence the previous lectotypification is to be disregarded. Accordingly, a new lectotype is designated for that name, currently accepted as Verbesina encelioides.