iWildCam 2021
Overview
Camera Traps (or Wild Cams) enable the automatic collection of large quantities of image data. Biologists all over the world use camera traps to monitor biodiversity and population density of animal species. We have recently been making strides towards automatic species classification in camera trap images, and in iWildCam 2020 we focused on generalization to novel camera locations across the globe. However, in order to estimate the abundance and density of species in camera trap data, biologists need to know how many of each species was seen. Because images are taken in bursts, object detection alone is not sufficient as it might lead to over- or under-counting, for example, if you get 3 images taken at one frame per second, and in the first you see 3 gazelles, in the second you see 5 gazelles, and in the last you see 4 gazelles, how many total gazelles have you seen? This is more challenging than strictly detecting and categorizing species, as it requires reasoning and tracking of individuals across sparse temporal samples.
Competition
Start Date - 10 March 2021
End Date - 28th May 2021
Kaggle URL - https://www.kaggle.com/c/iwildcam2021-fgvc8
Github URL - https://github.com/visipedia/iwildcam_comp
Organizers
Sara Beery (Caltech), Arushi Agarwal (Caltech), Elijah Cole (Caltech), and Vighnesh Birodkar (Google)
Acknowledgements
We would like to thank Microsoft AI for Earth for hosting the data and providing compute via Azure, the Wildlife Conservation Society for providing the dataset and species labels, and CentaurLabs for providing count labels on our test set.