Study wild animals with camera data
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 automating the species classification challenge in camera traps, but as we try to expand the scope of these models from specific regions where we have collected training data to nearby areas we are faced with an interesting probem: how do you classify a species in a new region that you may not have seen in previous training data?In order to tackle this problem, we have prepared a challenge where the training data and test data are from different regions, namely The American Southwest and the American Northwest. The species seen in each region overlap, but are not identical, and the challenge is to classify the test species correctly. To this end, we will allow training on our American Southwest data (from CaltechCameraTraps), on iNaturalist 2017/2018 data, and on simulated data generated from Microsoft AirSim. We have provided a taxonomy file mapping our classes into the iNat taxonomy.This is an FGVCx competition as part of the FGVC^6 workshop at CVPR 2019. Please open an issue if you have questions or problems with the dataset.There is $2,000 sponsored by Microsoft AI for Earth that will be awarded at the FGVC^6 workshop to top challenge solutions at the discretion of the competition hosts. At the end of the competition we will invite the top-scoring teams to document and open-source their methods and models. The prize money will be awarded to one or more teams based on their solutions: a combination of competition results, documentation and open-sourcing of methods and models, and method novelty.