Fine-tunes EG3D using a human-preference reward on NeRF density to improve face geometry, achieving 74.4% user preference in pairwise tests with FID rising from 4.09 to 6.66.
hub Mixed citations
Structured 3D Latents for Scalable and Versatile 3D Generation
Mixed citation behavior. Most common role is background (40%).
62
Pith papers citing it
6
external citations · Pith
Background
40% of classified citations
abstract
We introduce a novel 3D generation method for versatile and high-quality 3D asset creation. The cornerstone is a unified Structured LATent (SLAT) representation which allows decoding to different output formats, such as Radiance Fields, 3D Gaussians, and meshes. This is achieved by integrating a sparsely-populated 3D grid with dense multiview visual features extracted from a powerful vision foundation model, comprehensively capturing both structural (geometry) and textural (appearance) information while maintaining flexibility during decoding. We employ rectified flow transformers tailored for SLAT as our 3D generation models and train models with up to 2 billion parameters on a large 3D asset dataset of 500K diverse objects. Our model generates high-quality results with text or image conditions, significantly surpassing existing methods, including recent ones at similar scales. We showcase flexible output format selection and local 3D editing capabilities which were not offered by previous models. Code, model, and data will be released.
hub tools
citation-role summary
background 4
baseline 2
dataset 2
method 2
citation-polarity summary
representative citing papers
Arbor attaches constraint mesh tokens to a frozen text-to-3D denoiser to enable controllable generation obeying hull, avoidance, and touch constraints.
MeshFlow applies equivariant optimal-transport flow matching to generate triangle meshes as soups, matching autoregressive quality with an 18x inference speedup.
AdaVoMP predicts accurate dense spatially-varying Young's modulus, Poisson's ratio and density for 3D objects using an adaptive sparse voxel structure generated by a sparse transformer encoder-decoder at 16^3 higher resolution than prior fixed-voxel methods.
Garment Particles is a 5D point cloud representation jointly encoding 2D sewing patterns and 3D geometry, supporting rectified flow generation from high-level inputs and diffusion-based editing of patterns or shapes.
GenRecon lifts object-level generative priors to scene-scale reconstruction by chunking scenes and using projection-based conditioning on multi-view features, claiming 16% better results than prior methods.
CAdam reinterprets densification in generative 3DGS as signal verification via gradient-moment interference, quantile context, and SNR gating to achieve large reductions in primitive count with comparable quality.
A video generation approach conditions a base model with multi-scale 3D latent features and a cross-attention adapter to produce geometrically realistic and consistent orbital videos from one image.
SEM-ROVER generates large multiview-consistent 3D urban driving scenes via semantic-conditioned diffusion on Σ-Voxfield voxel grids with progressive outpainting and deferred rendering.
MeshTailor is a mesh-native generative model that uses ChainingSeams serialization and a dual-stream transformer with pointer layers to trace coherent seams vertex-by-vertex on 3D surfaces.
ATATA enables fast joint inference of structurally aligned pairs using Rectified Flow models via segment transport, improving state-of-the-art for image and video generation while matching 3D quality at much higher speed.
Affostruction reconstructs full 3D object geometry from partial RGBD views and grounds text-based affordances on both visible and unobserved surfaces, reporting large gains over prior methods.
Voxify3D generates voxel art from 3D meshes via orthographic pixel supervision, patch-based CLIP alignment, and palette-constrained Gumbel-Softmax quantization, achieving 37.12 CLIP-IQA and 77.90% user preference.
SVG360 lifts a single SVG to a view-conditioned representation, uses spatial memory to propagate consistent parts across views, and applies structure-aware vectorization to produce editable multiview SVGs.
GenHSI is a training-free three-stage pipeline that turns a scene image, character image, and complex HSI prompt into long videos with plausible chained interactions by generating atomic actions, 3D keyframes via 2D inpainting plus optimization, and then feeding them to pre-trained video diffusion.
An end-to-end 3D editing framework achieves high-fidelity local edits from coarse bounding boxes and 2D image prompts using region-aware loss reweighting and a large-scale parts-derived training dataset.
A single RGB image is converted into a layered, simulation-ready robot scene that supports trajectory replay, synthetic data generation, and meaningful sim-real policy evaluation, plus a 564-scene DROID-Sim companion set.
SynCity 3000 generates large, coherent 3D scenes from text by fine-tuning an image-to-3D diffusion model to operate convolutionally on overlapping windows, trained on procedurally generated synthetic scene data.
PixGS is a single-stage pixel-space diffusion model that directly produces high-quality 3D Gaussian Splats from text or images in ~1s, outperforming multi-stage latent methods on standard benchmarks.
SubdivAR reformulates neural mesh subdivision as autoregressive next-scale vertex-offset prediction, reporting 18.8% lower Hausdorff and 14.2% lower Chamfer distance than NMR on closed meshes.
HiFiVe is a training-free framework using an auto-regressive texture refinement pipeline with depth-based warping, multi-view fusion, and symmetry to enhance both texture and geometry fidelity in vehicle generation from 2D priors.
Diffusion-based per-view harmonization for lighting-consistent object transfer between 3DGS scenes, using heterogeneous training data and final 3D consolidation.
A de-biased VLM judge protocol is applied to adapt TRELLIS for single-image furniture 3D generation but yields no improvement over the strong public base across six methods.
DO AS I DO reconstructs and retargets hand-object interactions from in-the-wild monocular RGB videos to produce dexterous robot manipulation trajectories, outperforming prior methods on ground-truth and online video datasets.
citing papers explorer
-
Sculpting NeRF Geometry: Human-Preference Fine-Tuning of a 3D-Aware Face GAN
Fine-tunes EG3D using a human-preference reward on NeRF density to improve face geometry, achieving 74.4% user preference in pairwise tests with FID rising from 4.09 to 6.66.
-
Arbor: Explicit Geometric Conditioning for Controllable 3D Asset Generation
Arbor attaches constraint mesh tokens to a frozen text-to-3D denoiser to enable controllable generation obeying hull, avoidance, and touch constraints.
-
MeshFlow: Mesh Generation with Equivariant Flow Matching
MeshFlow applies equivariant optimal-transport flow matching to generate triangle meshes as soups, matching autoregressive quality with an 18x inference speedup.
-
Adaptive Volumetric Mechanical Property Fields Invariant to Resolution
AdaVoMP predicts accurate dense spatially-varying Young's modulus, Poisson's ratio and density for 3D objects using an adaptive sparse voxel structure generated by a sparse transformer encoder-decoder at 16^3 higher resolution than prior fixed-voxel methods.
-
Garment Particles: A 2D--3D Symmetric Garment Representation for Generation and Editing
Garment Particles is a 5D point cloud representation jointly encoding 2D sewing patterns and 3D geometry, supporting rectified flow generation from high-level inputs and diffusion-based editing of patterns or shapes.
-
GenRecon: Bridging Generative Priors for Multi-View 3D Scene Reconstruction
GenRecon lifts object-level generative priors to scene-scale reconstruction by chunking scenes and using projection-based conditioning on multi-view features, claiming 16% better results than prior methods.
-
CAdam: Context-Adaptive Moment Estimation for 3D Gaussian Densification in Generative Distillation
CAdam reinterprets densification in generative 3DGS as signal verification via gradient-moment interference, quantile context, and SNR gating to achieve large reductions in primitive count with comparable quality.
-
Towards Realistic and Consistent Orbital Video Generation via 3D Foundation Priors
A video generation approach conditions a base model with multi-scale 3D latent features and a cross-attention adapter to produce geometrically realistic and consistent orbital videos from one image.
-
SEM-ROVER: Semantic Voxel-Guided Diffusion for Large-Scale Driving Scene Generation
SEM-ROVER generates large multiview-consistent 3D urban driving scenes via semantic-conditioned diffusion on Σ-Voxfield voxel grids with progressive outpainting and deferred rendering.
-
MeshTailor: Cutting Seams via Generative Mesh Traversal
MeshTailor is a mesh-native generative model that uses ChainingSeams serialization and a dual-stream transformer with pointer layers to trace coherent seams vertex-by-vertex on 3D surfaces.
-
ATATA: One Algorithm to Align Them All
ATATA enables fast joint inference of structurally aligned pairs using Rectified Flow models via segment transport, improving state-of-the-art for image and video generation while matching 3D quality at much higher speed.
-
Affostruction: 3D Affordance Grounding with Generative Reconstruction
Affostruction reconstructs full 3D object geometry from partial RGBD views and grounds text-based affordances on both visible and unobserved surfaces, reporting large gains over prior methods.
-
Voxify3D: Pixel Art Meets Volumetric Rendering
Voxify3D generates voxel art from 3D meshes via orthographic pixel supervision, patch-based CLIP alignment, and palette-constrained Gumbel-Softmax quantization, achieving 37.12 CLIP-IQA and 77.90% user preference.
-
SVG360: Editable Multiview Vector Graphics from a Single SVG
SVG360 lifts a single SVG to a view-conditioned representation, uses spatial memory to propagate consistent parts across views, and applies structure-aware vectorization to produce editable multiview SVGs.
-
GenHSI: Controllable Generation of Human-Scene Interaction Videos
GenHSI is a training-free three-stage pipeline that turns a scene image, character image, and complex HSI prompt into long videos with plausible chained interactions by generating atomic actions, 3D keyframes via 2D inpainting plus optimization, and then feeding them to pre-trained video diffusion.
-
EditVerse3D: High-Quality 3D Object Editing with Region-Aware Learning
An end-to-end 3D editing framework achieves high-fidelity local edits from coarse bounding boxes and 2D image prompts using region-aware loss reweighting and a large-scale parts-derived training dataset.
-
RoboSnap: One-Shot Real-to-Sim Scene Generation for Generalizable Robot Learning and Evaluation
A single RGB image is converted into a layered, simulation-ready robot scene that supports trajectory replay, synthetic data generation, and meaningful sim-real policy evaluation, plus a 564-scene DROID-Sim companion set.
-
SynCity 3000: Bootstrapping Scene-Scale 3D Diffusion
SynCity 3000 generates large, coherent 3D scenes from text by fine-tuning an image-to-3D diffusion model to operate convolutionally on overlapping windows, trained on procedurally generated synthetic scene data.
-
PixGS: Pixel-Space Diffusion for Direct 3D Gaussian Splat Generation
PixGS is a single-stage pixel-space diffusion model that directly produces high-quality 3D Gaussian Splats from text or images in ~1s, outperforming multi-stage latent methods on standard benchmarks.
-
SubdivAR: Autoregressive Next-Scale Prediction for Neural Mesh Subdivision
SubdivAR reformulates neural mesh subdivision as autoregressive next-scale vertex-offset prediction, reporting 18.8% lower Hausdorff and 14.2% lower Chamfer distance than NMR on closed meshes.
-
HiFiVe: High-Fidelity Vehicle Generation Leveraging Auto-Regressive 2D Generative Priors
HiFiVe is a training-free framework using an auto-regressive texture refinement pipeline with depth-based warping, multi-view fusion, and symmetry to enhance both texture and geometry fidelity in vehicle generation from 2D priors.
-
Lighting-Consistent Object Transfer Across Radiance Fields
Diffusion-based per-view harmonization for lighting-consistent object transfer between 3DGS scenes, using heterogeneous training data and final 3D consolidation.
-
Judging to Improve: A De-biased VLM-as-3D-Judge Protocol for Single-Image 3D Generation
A de-biased VLM judge protocol is applied to adapt TRELLIS for single-image furniture 3D generation but yields no improvement over the strong public base across six methods.
-
Do as I Do: Dexterous Manipulation Data from Everyday Human Videos
DO AS I DO reconstructs and retargets hand-object interactions from in-the-wild monocular RGB videos to produce dexterous robot manipulation trajectories, outperforming prior methods on ground-truth and online video datasets.
-
Surflo: Consistent 3D Surface Flow Model with Global State
Surflo compresses unposed RGB views into K global latent tokens and uses flow matching with photometric guidance to decode consistent arbitrary-resolution 3D surface points in one forward pass.
-
MeshFlow: Efficient Artistic Mesh Generation via MeshVAE and Flow-based Diffusion Transformer
MeshFlow uses a contrastive MeshVAE for compact mesh latents and a flow transformer for parallel generation, claiming 18x speedup over autoregressive methods with high accuracy on standard metrics.
-
PerceptTwin: Semantic Scene Reconstruction for Iterative LLM Planning and Verification
PerceptTwin creates interactive simulations from open-vocabulary object maps for verifying and refining LLM robot plans, reporting ~39% higher success rates and up to 18% better human verification.
-
PhyGenHOI: Physically-Aware 4D Generation of Dynamic Human-Object Interactions
PhyGenHOI couples a motion diffusion model for humans with material point method simulation for objects on 3D Gaussians, using attraction loss, contact re-simulation, and masked video-SDS to produce physically consistent dynamic interactions from text.
-
SuperVoxelGPT: Adaptive and Ordered 3D Tokenization for Autoregressive Shape Generation
Adaptive saliency-guided supervoxel tokenization cuts 3D AR token length to 12.8% of uniform voxels while claiming SOTA quality and ~10× speedup on Trellis-500K.
-
Fishbone: From One 3D Asset to a Million Controllable Edits
Fishbone introduces a unified rib-spine representation computed via adaptive heat method, iso-contour ribs, and geometry-aware spine that enables real-time parametric deformation, reduced-space simulation, and animation on general meshes.
-
PhysX-Omni: Unified Simulation-Ready Physical 3D Generation for Rigid, Deformable, and Articulated Objects
PhysX-Omni unifies simulation-ready 3D asset generation across rigid, deformable, and articulated objects via a new geometry representation, the PhysXVerse dataset, and the PhysX-Bench evaluation suite.
-
ROAR-3D: Routing Arbitrary Views for High-Fidelity 3D Generation
ROAR-3D adds a token-wise view router and dual-stream attention to pretrained single-view 3D generators so they can use arbitrary unposed images for higher-fidelity output.
-
Syn4D: A Multiview Synthetic 4D Dataset
Syn4D supplies multiview synthetic dynamic scenes with dense geometric, tracking and pose ground truth that lets any pixel be unprojected to any time and camera.
-
PhysForge: Generating Physics-Grounded 3D Assets for Interactive Virtual World
PhysForge generates physics-grounded 3D assets via a VLM-planned Hierarchical Physical Blueprint and a KineVoxel Injection diffusion model, backed by the new PhysDB dataset of 150,000 annotated assets.
-
Velox: Learning Representations of 4D Geometry and Appearance
Velox compresses dynamic point clouds into latent tokens that support geometry via 4D surface modeling and appearance via 3D Gaussians, showing strong results on video-to-4D generation, tracking, and image-to-4D cloth simulation.
-
MeshReGen: A Unified 3D Geometry Regeneration Framework
MeshReGen introduces a conditioned 3D geometry regenerator with VecSet that learns a regeneration prior via self-supervision and reports state-of-the-art results on controllable generation tasks.
-
REVIVE 3D: Refinement via Encoded Voluminous Inflated prior for Volume Enhancement
REVIVE 3D generates voluminous 3D assets from flat 2D images via an inflated prior construction followed by latent-space refinement, plus new metrics for volume and flatness validated by user study.
-
Pair2Scene: Learning Local Object Relations for Procedural Scene Generation
Pair2Scene generates complex 3D scenes beyond training data by training a network on local object-pair placement rules and applying them recursively with collision-aware sampling.
-
WARPED: Wrist-Aligned Rendering for Robot Policy Learning from Egocentric Human Demonstrations
WARPED synthesizes realistic wrist-view observations from monocular egocentric human videos via foundation models, hand-object tracking, retargeting, and Gaussian Splatting to train visuomotor policies that match teleoperation success rates on five tabletop tasks with 5-8x less collection effort.
-
ReplicateAnyScene: Zero-Shot Video-to-3D Composition via Textual-Visual-Spatial Alignment
ReplicateAnyScene performs fully automated zero-shot video-to-compositional-3D reconstruction by cascading alignments of generic priors from vision foundation models across textual, visual, and spatial dimensions.
-
Physically Grounded 3D Generative Reconstruction under Hand Occlusion using Proprioception and Multi-Contact Touch
Proprioception and multi-contact touch, fused with a physics-guided conditional diffusion model over a Structure-VAE SDF latent space, improve metric amodal object reconstruction under severe hand occlusion versus vision-only baselines.
-
UniRecGen: Unifying Multi-View 3D Reconstruction and Generation
UniRecGen unifies reconstruction and generation via shared canonical space and disentangled cooperative learning to produce complete, consistent 3D models from sparse views.
-
MV-SAM3D: Adaptive Multi-View Fusion for Layout-Aware 3D Generation
MV-SAM3D adds multi-view fusion via multi-diffusion with attention-entropy and visibility weighting plus physics-aware optimization to improve fidelity and physical plausibility in layout-aware 3D generation.
-
Depth Anything 3: Recovering the Visual Space from Any Views
DA3 recovers consistent visual geometry from arbitrary views via a vanilla DINO transformer and depth-ray target, setting new SOTA on a visual geometry benchmark while outperforming DA2 on monocular depth.
-
Art3D: Training-Free 3D Generation from Flat-Colored Illustration
Art3D enhances flat-colored 2D illustrations with 3D illusion using pre-trained 2D model features and VLM realism evaluation, then generates 3D, while introducing the Flat-2D benchmark dataset.
-
Restore3D: Breathing Life into Broken Objects with Shape and Texture Restoration
Restore3D restores shape and texture of broken 3D objects via multi-view image refinement with a Mask Self-Perceiver and coarse-to-fine mesh reconstruction, outperforming baselines on synthetic and real benchmarks.
-
JacobianAvatar: Temporally Consistent Semi-rigid Avatar Reconstruction from a Monocular Video
JacobianAvatar uses neural Jacobian fields with a constrained Poisson solver, signed distance regularization, and deformation-guided flow loss to produce temporally consistent avatars from monocular video.
-
ReScene: Structured Indoor Scene Reconstruction from Multi-View Captures
ReScene introduces HierView for view prioritization and Relation-Aware Assembly for scene graph fusion, reporting 17% lower Chamfer Distance and 26% lower LPIPS than prior baselines on ScanNet while running faster.
-
Generative Relightable Avatars
GRA combines a microfacet-relit 3D avatar with a fine-tuned video diffusion model to generate photorealistic, freely viewable relit videos of a person under arbitrary environment lighting.
-
GASE: Gaussian Splatting-Based Automated System for Reconstructing Embodied-Simulation Environments
GASE automates high-fidelity simulation scene reconstruction from multi-view panoramic videos via Gaussian splatting, object extraction, and inpainting, yielding robot policies with under 10% performance gap versus real-world training.