New Publication: “PolarBearVidID: A Video-Based Re-Identification Benchmark Dataset for Polar Bears”

Symbolic picture for the article. The link opens the image in a large view.
The MaD Lab’s Animal Behavior Research Group Team VERA published their latest dataset 🎉🐻‍❄️
Zuerl, M.; Dirauf, R.; Koeferl, F.; Steinlein, N.; Sueskind, J.; Zanca, D.; Brehm, I.; Fersen, L.v.; Eskofier, B. PolarBearVidID: A Video-Based Re-Identification Benchmark Dataset for Polar Bears. Animals 202313, 801.
Simple Summary:
Zoos use automated systems to study animal behavior. These systems need to be able to identify animals from different cameras. This can be challenging, as individuals of the same species might look very alike. Deep Learning is the best way to automatically perform this task, especially when using videos instead of images because they show the animal’s movement as additional information. To train the Deep Learning model, one needs to have data. This study introduces a new dataset called PolarBearVidID that includes video sequences of 13 polar bears in various poses and lighting conditions. Our model is able to identify them with 96.6% accuracy. This shows that using the animals’ movements can help identify them.