iWildCam 2018

Camera Traps (or Wild Cams) enable the automatic collection of large quantities of image data. Unfortunately, a large number of the images that end up being captured tend to be false positives. This is typically caused by non-animal motion in the scene e.g. the wind moving trees.

The goal of this competition is to predict if images from a diverse set of unseen locations that have been captured both during the day and at night contain any animals. The main challenge is generalizing to camera trap locations that are not present in the training set. Another challenge is that some images contain other objects (e.g. people or vehicles) that can trigger the cameras but are not of interest.

Checkout the iWildCam Competition Github repo for the specifics of the dataset and download links. The competition itself is being hosted on Kaggle.

Competition Begins March 2018

Submission Deadline 4th June 2018

Data is primarily provided by Erin Boydston (USGS) and Justin Brown (NPS).