@article{fdi:010079836,
  title = {{D}eep{F}orest : {A} {P}ython package for {RGB} deep learning tree crown delineation},
  author = {{W}einstein, {B}. and {M}arconi, {S}. and {A}ubry-{K}ientz, {M}. and {V}incent, {G}r{\'e}goire and {S}enyondo, {H}. and {W}hite, {E}. {P}.},
  editor = {},
  language = {{ENG}},
  abstract = {{R}emote sensing of forested landscapes can transform the speed, scale and cost of forest research. {T}he delineation of individual trees in remote sensing images is an essential task in forest analysis. {H}ere we introduce a new {P}ython package, {D}eep{F}orest that detects individual trees in high resolution {RGB} imagery using deep learning. {W}hile deep learning has proven highly effective in a range of computer vision tasks, it requires large amounts of training data that are typically difficult to obtain in ecological studies. {D}eep{F}orest overcomes this limitation by including a model pretrained on over 30 million algorithmically generated crowns from 22 forests and fine-tuned using 10,000 hand-labelled crowns from six forests. {T}he package supports the application of this general model to new data, fine tuning the model to new datasets with user labelled crowns, training new models and evaluating model predictions. {T}his simplifies the process of using and retraining deep learning models for a range of forests, sensors and spatial resolutions. {W}e illustrate the workflow of {D}eep{F}orest using data from the {N}ational {E}cological {O}bservatory {N}etwork, a tropical forest in {F}rench {G}uiana, and street trees from {P}ortland, {O}regon.},
  keywords = {crown delineation ; deep learning ; forests ; {NEON} ; remote sensing ; {RGB} ; tree crowns},
  booktitle = {},
  journal = {{M}ethods in {E}cology and {E}volution},
  volume = {11},
  numero = {12},
  pages = {1743--1751},
  ISSN = {2041-210{X}},
  year = {2020},
  DOI = {10.1111/2041-210x.13472},
  URL = {https://www.documentation.ird.fr/hor/fdi:010079836},
}