Enlarge the image

DeepFaune, automated image processing software for photographic traps in France

The DeepFaune initiative, led by members of LBBE (Vincent Miele, Bruno Spataro and Gaspard Dussert) and CEFE, as well as a network of over 40 biodiversity stakeholders, is presented in the technical journal of the French Office for Biodiversity:

https://www.ofb.gouv.fr/sites/default/files/Fichiers/Plaquettes%20et%20rapports%20instit/biodiversite4-deepfaune.pdf

Published on

Led by members of LBBE (Vincent Miele, Bruno Spataro and Gaspard Dussert) and CEFE (Simon Chamaillé-Jammes, Noa Rigoudy), the DeepFaune initiative aims to develop a free, open-source tool to help sort animal images and videos from camera traps, using artificial intelligence.

From the outset, the DeepFaune initiative was designed to be collaborative. A wide range of partners (OFB, national and regional parks, hunting federations, associations, research teams, etc.) have been asked to support the initiative, by formulating their needs and contributing to the image bank needed to train the artificial intelligence model. With more than a million images currently annotated, this fruitful collaboration provides highly favourable conditions for the development of a high-performance recognition model.

The latest version of the software, version 1.0.0, has been extensively modified and improved, and is available for free download from https://www.deepfaune.cnrs.fr/.