iWildCam2022

Overview

Camera Traps (or Wild Cams) enable the automatic collection of large quantities of image data. Biologists all over the world use them to monitor biodiversity and population density of animal species. In recent years, we have made good progress on automatic species classification in camera trap images and generalization to novel camera locations across the globe.

However, in order to estimate the abundance and density of species in camera trap data, biologists also need to know how many of each species were 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?

In iWildCam 2021 we combined species classification with counting of individual animals across sparse temporal samples and we learned about the difficulties on solving and interpreting the results on two hard tasks at the same time. This year, instead, we would like to focus entirely on the counting task, since reasoning across image sequences poses interesting challenges in itself that deserve more attention from our community.

Competition

Start Date: 21 March 2022

End Date: 30 May 2022

Kaggle URL: https://www.kaggle.com/c/iwildcam2022-fgvc9/

Github URL: https://github.com/visipedia/iwildcam_comp

Organizers

• Stefan Istrate (Tech4Nature)

• Sara Beery (Caltech)

• John Beuving