Research system that turns a single image into a 360° Gaussian scaffold and renders interactive video at 8 FPS.
Updated EU status checked Last shown
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
MoVerse is a research project from the Orange Team (Youku Moku-Lab, HUJING Digital Media & Entertainment Group) on real-time video world modeling. Given one narrow-field-of-view image, it builds a navigable 3D world and renders video along a user-specified camera trajectory.
The pipeline has three stages. Stage I expands the input image into a gravity-aligned, horizontally periodic 360° panorama using topology-aware latent diffusion. Stage II lifts the panorama into a panoramic 3D Gaussian scaffold via feed-forward residual prediction in angular–inverse-depth space. Stage III is a causal renderer that translates scaffold renderings into photorealistic video, reportedly at 8 FPS on a single RTX 4090. The scaffold is reusable and keeps geometry consistent when the camera revisits areas.
Currently only a project page with demos is available. A YouTube reviewer reported that output quality is somewhat blurry and that code release is pending a compliance review. No weights, license, or paper link were found in the provided material.
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Converts one image into a 360° panorama then 3D Gaussian world, rendering interactive video at 8 fps on an RTX 4090. Quality is blurry and code is pending compliance review.
Closest neighbours by meaning, not just by category.
Model
Research framework that turns posed images with instance masks and 3D boxes into per-object meshes for cluttered scenes.
Model · Z.ai
Research framework for end-to-end controlled character animation and motion transfer from a driving video to a reference character.
Model
3D-aware 360° video diffusion model that generates panoramic videos along user-defined camera paths from sparse panoramas.
Research
Research method that reconstructs complete 4D (3D plus time) scenes from a single monocular video using Gaussian Splatting.
Research · ByteDance
ByteDance research pipeline that turns indoor captures into editable, simulation-ready 3D scenes of separate object meshes.
Model
World model that turns a few context images into an editable, physics-simulated 3D world rendered by a video model.
