=Paper=
{{Paper
|id=Vol-1795/paper45
|storemode=property
|title=FLOPO - An Ontology for the Integration of Trait Data from Digitized Floras and Plant Image Collections
|pdfUrl=https://ceur-ws.org/Vol-1795/paper45.pdf
|volume=Vol-1795
|authors=Claus Weiland,Robert Hoehndorf,George Gosline,Quentin Groom,Thomas Hamann,Marco Schmidt
|dblpUrl=https://dblp.org/rec/conf/swat4ls/WeilandHGGHS16
}}
==FLOPO - An Ontology for the Integration of Trait Data from Digitized Floras and Plant Image Collections==
<pdf width="1500px">https://ceur-ws.org/Vol-1795/paper45.pdf</pdf>
<pre>
   FLOPO - An Ontology for the Integration of
    Trait Data from Digitized Floras and Plant
                 Image Collections

   Claus Weiland1 , Robert Hoehndorf2 , George Gosline3 , Quentin Groom4 ,
                   Thomas Hamann5 , and Marco Schmidt1
        1
            Senckenberg Biodiversity and Climate Research Centre Frankfurt
            2
              King Abdullah University of Science and Technology, Thuwal
                          3
                            Royal Botanical Gardens Kew
                              4
                                Botanic Garden Meise
                       5
                         Naturalis Biodiversity Center Leiden

    Knowledge about organismic traits is useful for many different studies in
environmental sciences. For example, traits can be used to study the functional
diversity of plant communities and competition between species. Furthermore
traits of underutilized plants are an important resource for agricultural usage and
phytomedical applications. There are large biodiversity data collections which
are a potentially rich source of such plant trait data:
 – Floras describe plant traits with a focus on morphology or habitus relevant
   for species identification in addition to other characteristics (e.g. health ap-
   plications).
 – Complementary topic databases, for example focusing on geographical dis-
   tributions like Senckenberg’s West African Plants [1], provide large image
   collections which allow additional trait and habitat data to be extracted.
However, two key limitations in systematically analyzing trait data are the lack
of a standardized vocabulary for the described traits and difficulties in extracting
structured information from different knowledge sources. We have developed the
Flora Phenotype Ontology (FLOPO [2]), an ontology for describing traits of
plant species found in Floras.
    Our methodological approach was to use NLP to extract entity-quality rela-
tionships from digitized taxon descriptions in Floras [3]. On this basis, we used
a formal approach based on phenotype description patterns and automated rea-
soning to generate the FLOPO, which comprises more than 24000 phenotype
and trait classes. We are currently applying deep learning approaches (convolu-
tional neural networks) to extract additional plant trait data from image data
like herbarium scans and photographs.


References
1. http://www.westafricanplants.senckenberg.de
2. http://aber-owl.net/ontology/FLOPO
3. Hoehndorf, R. et al.: Journal of Biomedical Semantics 7:65 (2016)

</pre>