EcoTaxa is a web application destined to process the large number of images generated by such quantitative imaging instruments. It leverages deep learning and an efficient user interface to allow taxonomists to classify thousands of images per day. In addition, it can store a large quantity of metadata together with the images, with very few constraints on its content. These metadata can be used by the operators to sort through the images and by the machine learning backend to suggest identifications. Finally, EcoTaxa can export the metadata and the identifications together, in a versatile text table or following the DarwinCore Archive standard, for further data exploitation.
This is the front-end part of the application, which is (nearly) stateless and takes/writes all data from/to a back-end.
To run a simple set up see the all_in_one instructions. This will give you a running solution but probably not scale to be a production environment.
See the backend's instructions to setup a development environment.
To start, provided that you have a proper running back-end, you must create a short file config/config.cfg
. You can use appli/config-model.cfg
as a template.
Then read the documentation of this repository to
This project is tested with BrowserStack.
If you use EcoTaxa in your work, please cite it as
Marc Picheral, Sébastien Colin, and Jean-Olivier Irisson (2017) EcoTaxa, a tool for the taxonomic classification of images. http://ecotaxa.obs-vlfr.fr
This code is released under the GPLv3.
Specifications and supervision
Current developers
Past contributors