Logo Lanfrica

Overview of BirdCLEF 2023: Automated Bird Species Identification in Eastern Africa

Domaine:

environment and energy

Type de record:

paperdataset
Créateur:
KahDenKliRee
Éditeur:
CheCorGooÖko
Éditeur:
CCSD
Hôte:avatar
Source Agritrop Cirad (agritrop.cirad.fr) International audience The BirdCLEF 2023 challenge focused on bird species classification in a dataset of Kenyan soundscape recordings. Kenya is home to over 1,000 species of birds, covering a wide range of ecosystems, from the savannahs of the Maasai Mara to the Kakamega rainforest, and even alpine regions on Kilimanjaro and Mount Kenya. Tracking this vast number of species with ML can be challenging, especially with minimal training data available for many species. This year the competition switched back to threshold-free evaluation metric, and introduced a two-hour time limit on inference to ensure the practical usability of models.