The Intelligence Advanced Research Projects Activity, within the Office of the Director of National Intelligence, announced that they have selected the winners of the Multi-View Stereo 3D Mapping Challenge.
At an all-day workshop in Washington D.C. on November 30, 2016, the top solvers demoed their solutions with peers and other members of the IARPA community, sharing their novel approaches to developing Multi-View Stereo 3D mapping algorithms. Solutions that were open sourced were released during this event, and will be made available to the public.
These winners developed the best algorithms to accurately produce 3D mappings from satellite photos out of a total of 34 participants who responded to the challenge. The challenge was divided into two segments: the Explorer Challenge marked the first half of the challenge, which served as a way to introduce solvers to the data-set and to prepare for the Master Challenge. In the Master Challenge, solvers had to take their Explorer Challenge experience to design the strongest possible mapping algorithm.
The finalists have all made significant contributions to fostering innovation through crowd-sourcing and together they form a collaborative community that helps advance new and better methods for building 3D point clouds. IARPA thanks all solvers for their participation and looks forward to creating more innovative solutions with community collaboration in the future.
To learn more about the Multi-View Stereo 3D Mapping Challenge, including eligibility requirements, problem details, results, findings, and the full list of winners, click here.
Source: Press Release