4 papers indexed
Gal Fiebelman, Hadar Averbuch-Elor, Sagie Benaim
Recent advances in video diffusion models have enabled either long single-view generation through temporal autoregression, or short multi-view synthesis through bidirectional attention. However, generating long, multi-view consistent videos of dynamic scenes remains unsolved. In…
Nitzan Hodos, Roy Amoyal, Lior Fritz, Ianir Ideses, Sagie Benaim, Netalee Efrat
Close-up rendering, zooming into a scene well beyond any training camera, is important for virtual production and interactive 3D content, yet remains an open challenge. 3D Gaussian splatting (3DGS) enables high-fidelity, real-time novel view synthesis, but its rendering quality d…
Omer Sela, Inbar Huberman-Spiegelglas, Michael Rotman, Sagie Benaim, Avi Ben-Cohen
Controlling the motion of multiple objects in image-to-video (I2V) generation requires preserving object identities while enforcing adherence to distinct target trajectories. This becomes particularly challenging as the number of objects increases and their paths intersect or occ…
Or Hirschorn, Aaron Olender, Eli Alshan, Ianir Ideses, Lior Fritz, Sagie Benaim
We present a zero-shot, training-free and optimization-free framework for generating 360 panoramic images and videos by directly injecting spherical priors into pre-trained diffusion transformers. Existing methods either rely on costly fine-tuning on scarce panoramic data that li…