Résumé
Building accurate knowledge of the identity, the geographicdistribution and the evolution of living species is essentialfor a sustainable development of humanity as well as forbiodiversity conservation. In this context, using multimediaidentication tools is considered as one of the most promisingsolution to help bridging the taxonomic gap. With therecent advances in digital devices/equipment, network bandwidthand information storage capacities, the production ofmultimedia big data has indeed become an easy task. In parallel,the emergence of citizen sciences and social networkingtools has fostered the creation of large and structured communitiesof nature observers (e.g. eBird, Xeno-canto, TelaBotanica, etc.) that have started to produce outstandingcollections of multimedia records. Unfortunately, the performanceof the state-of-the-art multimedia analysis techniqueson such data is still not well understood and is far fromreaching the real world's requirements in terms of identi-cation tools. The LifeCLEF lab proposes to evaluate thesechallenges around 3 tasks related to multimedia informationretrieval and ne-grained classication problems in 3 livingworlds. Each task is based on large and real-world data andthe measured challenges are dened in collaboration withbiologists and environmental stakeholders in order to reflect realistic usage scenarios.