WildAI

Idealised visualization of a digital wildlife sandbox to study animal / environment interactions (From Tuia et al., 2022)

Mapping wildlife-environment interactions in the Swiss Alps with AI

In times when wildlife diversity is threatened by environmental changes, research in animal ecology and wildlife conservation is being accelerated and transformed by the recent advances in Machine Learning.

WildAI is an interdisciplinary project that aims to bring together behavioral neurosciences, ecology, environmental computational sciences, computer vision and machine learning to scale up conservation research. From unmanned aerial vehicles (UAV) and camera trap (CT) acquired images, we seek to model the wildlife habitats in the Swiss National Park to complement ecological knowledge of how wildlife interacts with its environment and how perturbations in the environment can in turn affect wildlife behavior. 

Technically, perspectives of the project touch upon automation of species classification, animal behavior synthesis, individual re-identification, habitat reconstruction and spatio-temporal modeling of multi-species & species-environment interactions.

Partners

Funding

iPhD: EPFL interfaculty program between ENAC and SV