@article{fdi:010075504,
  title = {{T}oward a large-scale and deep phenological stage annotation of herbarium specimens : case studies from temperate, tropical, and equatorial floras},
  author = {{L}orieul, {T}. and {P}earson, {K}. {D}. and {E}llwood, {E}. {R}. and {G}oeau, {H}. and {M}olino, {J}ean-{F}ran{\c{c}}ois and {S}weeney, {P}. {W}. and {Y}ost, {J}. {M}. and {S}achs, {J}. and {M}ata-{M}ontero, {E}. and {N}elson, {G}. and {S}oltis, {P}. {S}. and {B}onnet, {P}. and {J}oly, {A}.},
  editor = {},
  language = {{ENG}},
  abstract = {{P}remise of the {S}tudy {P}henological annotation models computed on large-scale herbarium data sets were developed and tested in this study. {M}ethods {H}erbarium specimens represent a significant resource with which to study plant phenology. {N}evertheless, phenological annotation of herbarium specimens is time-consuming, requires substantial human investment, and is difficult to mobilize at large taxonomic scales. {W}e created and evaluated new methods based on deep learning techniques to automate annotation of phenological stages and tested these methods on four herbarium data sets representing temperate, tropical, and equatorial {A}merican floras. {R}esults {D}eep learning allowed correct detection of fertile material with an accuracy of 96.3%. {A}ccuracy was slightly decreased for finer-scale information (84.3% for flower and 80.5% for fruit detection). {D}iscussion {T}he method described has the potential to allow fine-grained phenological annotation of herbarium specimens at large ecological scales. {D}eeper investigation regarding the taxonomic scalability of this approach is needed.},
  keywords = {convolutional neural network ; deep learning ; herbarium data ; natural history collections ; phenological stage annotation ; visual data ; classification ; {ETATS} {UNIS} ; {GUYANE} {FRANCAISE} ; {ZONE} {TEMPEREE} ; {ZONE} {TROPICALE} ; {ZONE} {EQUATORIALE}},
  booktitle = {},
  journal = {{A}pplications in {P}lant {S}ciences},
  volume = {7},
  numero = {3},
  pages = {e1233 [14 p.]},
  ISSN = {2168-0450},
  year = {2019},
  DOI = {10.1002/aps3.1233},
  URL = {https://www.documentation.ird.fr/hor/fdi:010075504},
}