Papers
No Pose, No Problem in 4D: Feed-Forward Dynamic Gaussians from Unposed Multi-View Videos
Matteo Balice, Yanik Kunzi, Chenyangguang Zhang, Matteo Matteucci, Marc Pollefeys, Sungwhan Hong
A feed-forward method for reconstructing dynamic 3D scenes as Gaussian splats from multi-view video, without requiring known camera poses. The model estimates both scene geometry and camera parameters jointly in a single forward pass.
Scaling Image Geo-Localization to Continent Level
Philipp Lindenberger, Paul-Edouard Sarlin, Jan Hosang, Matteo Balice, Marc Pollefeys, Simon Lynen, Eduard Trulls
A large-scale image geo-localization system that retrieves the location of a query image by matching it against a continent-scale database. The work focuses on scaling retrieval models to handle significantly larger and more diverse inputs than prior approaches.