Search papers, labs, and topics across Lattice.
7
0
6
A plug-and-play module, RESBev, fortifies BEV perception against sensor degradation and adversarial attacks by learning latent BEV state transitions, offering a practical route to more reliable autonomous driving systems.
Achieve stable, semantically-guided humanoid robot control with PhysiFlow, a framework that fuses vision-language-action with physics-aware whole-body control.
Stop neglecting your camera data: Fusion4CA shows how a few simple, plug-and-play modules can significantly boost 3D object detection performance by better exploiting RGB information in BEV fusion.
Imagine training autonomous vehicles in a blizzard, then switching to a sunny day, all within the same reconstructed 4D urban scene – WeatherCity makes this a reality with unprecedented control and fidelity.
By "dreaming" plausible scene completions, Dream-SLAM enables robots to navigate dynamic environments more effectively, achieving better localization, mapping, and exploration than existing methods.
Endoscopic SLAM gets a serious upgrade: deformation-aware 3D Gaussian Splatting slashes pose estimation error by up to 50% while producing higher-fidelity reconstructions.
Reconstructing surgical scenes from monocular endoscope videos with large camera motion just got a whole lot better, thanks to a new window-based approach that doesn't need stereo depth or perfect camera tracking.
