Carolina, Vish, and Madhu met in the summer of 2022. They were interested in the intersection between deep learning, computer vision, and animal conservation.

The Professor Alberto Todeschini shared with the team an initiative in collaboration with the BAIR group.

The initiative consisted in developing a model that could re-identify animals in protected areas using images from camera traps in the wild.

Today, leopard spotting can re-identify pride individuals using images from the wild. It also detects new pride members and flags them for future expert domain verification.

Our work can save researchers countless hours of segregating leopard images from the raw captures from the wild.

Leopard spotting detects and keeps track of the current leopard population, providing insights about abrupt changes.

The output data can facilitate rangers take management decisions in regard to modifying landscapes or hunting regulations.