Search papers, labs, and topics across Lattice.
University of Toronto
4
0
6
3
Achieving state-of-the-art 6-DoF pose tracking accuracy without the need for 3D models or depth maps, ProxyPose redefines the landscape of monocular video tracking.
Stop sacrificing either global structure or fine details when extrapolating DiTs to higher resolutions – SEGA dynamically adjusts attention based on spatial frequency, giving you both.
Unlock efficient 4D object understanding from dynamic point clouds with Velox, a representation that's descriptive, compressive, and accessible.
Your camera's AI could be subtly rewriting reality, but this method lets you reverse the changes and see the "unhallucinated" original.
