Research method that generates compositional 3D scenes from a single image via object-to-scene distillation.
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.
DistScene is a research project for 3D scene generation from a single input image. According to its project page, it uses an object-to-scene distillation approach to automatically produce scalable training data, and a unified scene representation in which scene frame generation is followed by object-centric refinement. The environment and individual objects are generated in a shared coordinate system, and each object is then refined.
The project page offers an interactive 3D preview (a compressed version, about 10% of the full generated result), rendered video demos, and a scene-to-navigation simulation showing that the generated compositional scenes can be imported into navigation simulators. A YouTube creator summary says it performs best on benchmarks and that the model and data are to be released; this is not confirmed on the project page.
The authors are Kunming Luo, Hongyu Yan, Ken Deng, Chengcheng Zhou, Tianyu Liu, Haipeng Li, Haibin Huang, Xuelong Li and Ping Tan (arXiv preprint, 2026). No code, weights, or hosted service were found in the provided material, so it is currently a research demonstration only.
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Generates an entire 3D scene from a single image, with environment and objects in a shared coordinate system, then refines each object. Best on benchmarks; model and data to be released.
Closest neighbours by meaning, not just by category.
Research
Research framework that generates collision-free, physically stable 3D indoor rooms from diffusion-sampled layouts.
Model
Image-to-3D generator from Tsinghua and Tencent ARC Lab that builds meshes pixel-aligned to the input view.
Model
Research model that unifies 3D scene generation and reconstruction in a single pixel-space diffusion model.
Model
Research model that decomposes a single image or scene mesh into object-level 3D meshes with consistent spatial layout.
Model · Hi3D
Open-source image-to-3D and text-to-3D generation model from Hi3D, formerly Sparc3D, trained on ~2M curated 3D assets.
Model
Model that reconstructs a single image into a 3D scene with separate, individually usable objects.