Role • Graphics for 3D moment Client • New York Times Graphics Desk
The New York Times Graphics desk and NYTimes R&D teams joined forces to model key moments of the World Cup using machine learning and 3D reconstruction. As FIFA does not allow for any news organization to capture video at the stadium (besides channels broadcasting the games live) our team decided to recreate key actions of the World Cup by 3D reconstructing the scene in 3D from photographs using ICON (Implicit Clothed humans Obtained from Normals). This allowed for a more natural way to experience motion and (inherently spatial) high level sports, compared to still photography. Team : Karthik Patanjali, Mark McKeague, Alexandre Deveaux, Malika Khurana & Or Fleisher.
The method we developed in creating 3D imagery through machine learning became compelling for the World Cup as the New York Times and other media outlets are only authorized to shoot still photography from a single vantage point on-site. The still photography medium however, is not best suited to deconstruct motion, position and gesture in space, which sports journalism benefits from.
Through this method, we were able to extrapolate spatial information from a single photograph, allowing a reader to see the game from different perspectives and gaining more insight into the reality of goal scoring.