6 papers
Masked Visual Actions for Unified World Modeling
Hadi Alzayer, Wenlong Huang, Haonan Chen +8
Video models absorb rich priors over how the visual world moves, interacts, and responds to contact, making them promising substrates for robotic world modeling. The central challe…
Choreographing a World of Dynamic Objects
Yanzhe Lyu, Chen Geng, Karthik Dharmarajan +4
Dynamic objects in our physical 4D (3D + time) world are constantly evolving, deforming, and interacting with other objects, leading to diverse 4D scene dynamics. In this paper, we…
Coupled Diffusion Sampling for Training-Free Multi-View Image Editing
Hadi Alzayer, Yunzhi Zhang, Chen Geng +2
We present an inference-time diffusion sampling method to perform multi-view consistent image editing using pre-trained 2D image editing models. These models can independently prod…
Magic Fixup: Streamlining Photo Editing by Watching Dynamic Videos
Hadi Alzayer, Zhihao Xia, Xuaner Zhang +3
We propose a generative model that, given a coarsely edited image, synthesizes a photorealistic output that follows the prescribed layout. Our method transfers fine details from th…
Generative Multiview Relighting for 3D Reconstruction under Extreme Illumination Variation
Hadi Alzayer, Philipp Henzler, Jonathan T. Barron +3
Reconstructing the geometry and appearance of objects from photographs taken in different environments is difficult as the illumination and therefore the object appearance vary acr…
SimVS: Simulating World Inconsistencies for Robust View Synthesis
Alex Trevithick, Roni Paiss, Philipp Henzler +9
Novel-view synthesis techniques achieve impressive results for static scenes but struggle when faced with the inconsistencies inherent to casual capture settings: varying illuminat…