xView Computer Vision/Machine Learning Challenge

The Defense Innovation Unit (DIU) is sponsoring its second annual xView Challenge! This year’s xView2 Challenge encourages applied research in the computer vision community by asking GeoAI/ML experts to develop algorithms and models that identify building damage in post-disaster satellite imagery. For the purposes of this challenge, xView2 has announced that it will release a publicly available dataset consisting of very high-resolution satellite imagery annotated with building locations and damage scores before and after natural disasters.

The Department of Defense (DoD) believes this challenge will support the advancement of AI imagery analysis, which will increase the effectiveness of homeland security procedures following natural disasters.

Click to learn more about the challenge if you are interested in developing and networking your GeoAI/ML skillset while contributing to the mission of U.S. Homeland Security.

Registration is officially open!

xView2 Challenge data was released on September 23rd and submissions are due by November 22, 2019. Winners will be announced December 13, 2019, at the Humanitarian Assistance and Disaster Recovery (HADR) workshop at NeurIPS. The competition offers compelling awards and recognition for winning submissions and as well as the opportunity for follow-on acquisition to further develop successful prototypes with the USG. Awards and eligibility are determined by technological merit and capacity to deliver proposed solutions.

Join the ranks of top academic institutions, teams, and researchers in AI, ML, Geospatial Intelligence, and GeoAI by registering for the xView2 Challenge using the following link: https://xview2.org/signup

We look forward to working with you!