Publications des agents du Cirad

Cirad

Overview of LifeCLEF 2023: Evaluation of AI models for the identification and prediction of birds, plants, snakes and fungi

Joly A., Botella C., Picek L., Kahl S., Goeau H., Deneu B., Marcos D., Estopinan J., Leblanc C., Larcher T., Chamidullin R., Sulc M., Hrúz M., Servajean M., Glotin H., Planqué R., Vellinga W.P., Klinck H., Denton T., Eggel I., Bonnet P., Müller H.. 2023. In : Arampatzis Avi (ed.), Kanoulas Evangelos (ed.), Tsikrika Theodora (ed.), Vrochidis Stefanos (ed.), Giachanou Anastasia (ed.), Li Dan (ed.), Aliannejadi Mohammad (ed.), Vlachos Michalis (ed.), Faggioli Guglielmo (ed.), Ferro Nicola (ed.). Experimental IR meets multilinguality, multimodality, and interaction: 14th International Conference of the CLEF Association, CLEF 2023, Thessaloniki, Greece, September 18–21, 2023, Proceedings. Cham : Springer, p. 416-439. (Lecture Notes in Computer Science, 14163). International Conference of the Cross-Language Evalution Forum for European Languages. 14, 2023-09-18/2023-09-21, Thessaloniki (Grèce).

Biodiversity monitoring through AI approaches is essential, as it enables the efficient analysis of vast amounts of data, providing comprehensive insights into species distribution and ecosystem health and aiding in informed conservation decisions. Species identification based on images and sounds, in particular, is invaluable for facilitating biodiversity monitoring efforts and enabling prompt conservation actions to protect threatened and endangered species. The LifeCLEF virtual lab has been promoting and evaluating advances in this domain since 2011. The 2023 edition proposes five data-oriented challenges related to the identification and prediction of biodiversity: (i) BirdCLEF: bird species recognition in long-term audio recordings (soundscapes), (ii) SnakeCLEF: snake identification in medically important scenarios, (iii) PlantCLEF: very large-scale plant identification, (iv) FungiCLEF: fungi recognition beyond 0–1 cost, and (v) GeoLifeCLEF: remote sensing-based prediction of species. This paper overviews the motivation, methodology, and main outcomes of that five challenges.

Documents associés

Communication de congrès

Agents Cirad, auteurs de cette publication :