Free-Range Gaussians uses flow matching over Gaussian parameters to predict non-grid-aligned 3D Gaussians from multi-view images, enabling synthesis of plausible content in unobserved regions with fewer primitives than grid-aligned methods.
hub Canonical reference
Meshlrm: Large reconstruction model for high- quality mesh
Canonical reference. 80% of citing Pith papers cite this work as background.
18
Pith papers citing it
Background
80% of classified citations
hub tools
citation-role summary
background 4
baseline 1
citation-polarity summary
fields
cs.CV 18representative citing papers
THOM is a training-free two-stage framework that generates physically plausible hand-object 3D meshes directly from text by combining text-guided Gaussians with contact-aware physics optimization and VLM refinement.
A latent-space transformer framework poses 3D characters without skinning or fixed topologies, outperforming baselines and generalizing zero-shot to quadrupeds.
ART is a category-agnostic transformer that maps sparse multi-state RGB images to per-part 3D geometry, texture, and articulation parameters via learnable part slots.
NeuMatEx extracts spatially varying neural materials from multi-view images by using a learned LMRM prior for initialization followed by uncertainty-guided inverse path tracing optimization.
Artic-O introduces an end-to-end feed-forward model that reconstructs articulated objects from sparse multi-state images via latent geometry learning and an image-grounded part-reasoning module, outperforming staged methods in speed on PartNet-Mobility while matching accuracy.
TriSplat predicts oriented triangle primitives from images in one forward pass to produce simulation-ready 3D meshes with competitive rendering quality.
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.
R-DMesh proposes a VAE-based disentanglement of base mesh, motion trajectories, and rectification offset plus Triflow Attention and rectified-flow diffusion to produce 4D meshes aligned to video despite initial pose mismatch.
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.
The paper proposes a problem-driven taxonomy for feed-forward 3D scene modeling that groups methods by five core challenges: feature enhancement, geometry awareness, model efficiency, augmentation strategies, and temporal-aware modeling.
Scaling transformer context with sparse attention and 3D-aware block routing improves feed-forward 3D reconstruction and inverse rendering, closing much of the quality gap with dense-view optimization.
Flux4D reconstructs large-scale dynamic 4D scenes unsupervised by predicting moving 3D Gaussians from photometric losses and static regularization when trained across multiple scenes.
TripoSG generates high-fidelity 3D meshes from input images via a large-scale rectified flow transformer and hybrid-trained 3D VAE on a custom 2-million-sample dataset, claiming state-of-the-art fidelity and generalization.
Decoder-only view synthesis model using KV-cache representation and weight sharing between reconstruction and rendering networks achieves new SOTA on novel view synthesis benchmarks.
KFC-W is a self-supervised 3D-aware video model trained on videos and multiview internet photos that produces geometrically consistent interpolations between unposed input images without any 3D annotations.
AnimateAnyMesh++ animates arbitrary 3D meshes from text using an expanded 300K-identity DyMesh-XL dataset, a power-law topology-aware DyMeshVAE-Flex, and a variable-length rectified-flow generator to produce semantically accurate, temporally coherent animations in seconds.
citing papers explorer
-
Free-Range Gaussians: Non-Grid-Aligned Generative 3D Gaussian Reconstruction
Free-Range Gaussians uses flow matching over Gaussian parameters to predict non-grid-aligned 3D Gaussians from multi-view images, enabling synthesis of plausible content in unobserved regions with fewer primitives than grid-aligned methods.
-
THOM: Generating Physically Plausible Hand-Object Meshes From Text
THOM is a training-free two-stage framework that generates physically plausible hand-object 3D meshes directly from text by combining text-guided Gaussians with contact-aware physics optimization and VLM refinement.
-
Make-It-Poseable: Feed-forward Latent Posing Model for 3D Characters
A latent-space transformer framework poses 3D characters without skinning or fixed topologies, outperforming baselines and generalizing zero-shot to quadrupeds.
-
ART: Articulated Reconstruction Transformer
ART is a category-agnostic transformer that maps sparse multi-state RGB images to per-part 3D geometry, texture, and articulation parameters via learnable part slots.
-
Extracting Neural Materials from Multi-view Images
NeuMatEx extracts spatially varying neural materials from multi-view images by using a learned LMRM prior for initialization followed by uncertainty-guided inverse path tracing optimization.
-
Artic-O: End-to-End Articulated Object Reconstruction via Latent Geometry Learning
Artic-O introduces an end-to-end feed-forward model that reconstructs articulated objects from sparse multi-state images via latent geometry learning and an image-grounded part-reasoning module, outperforming staged methods in speed on PartNet-Mobility while matching accuracy.
-
TriSplat: Simulation-Ready Feed-Forward 3D Scene Reconstruction
TriSplat predicts oriented triangle primitives from images in one forward pass to produce simulation-ready 3D meshes with competitive rendering quality.
-
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.
-
R-DMesh: Video-Guided 3D Animation via Rectified Dynamic Mesh Flow
R-DMesh proposes a VAE-based disentanglement of base mesh, motion trajectories, and rectification offset plus Triflow Attention and rectified-flow diffusion to produce 4D meshes aligned to video despite initial pose mismatch.
-
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.
-
Feed-Forward 3D Scene Modeling: A Problem-Driven Perspective
The paper proposes a problem-driven taxonomy for feed-forward 3D scene modeling that groups methods by five core challenges: feature enhancement, geometry awareness, model efficiency, augmentation strategies, and temporal-aware modeling.
-
LSRM: High-Fidelity Object-Centric Reconstruction via Scaled Context Windows
Scaling transformer context with sparse attention and 3D-aware block routing improves feed-forward 3D reconstruction and inverse rendering, closing much of the quality gap with dense-view optimization.
-
Flux4D: Flow-based Unsupervised 4D Reconstruction
Flux4D reconstructs large-scale dynamic 4D scenes unsupervised by predicting moving 3D Gaussians from photometric losses and static regularization when trained across multiple scenes.
-
TripoSG: High-Fidelity 3D Shape Synthesis using Large-Scale Rectified Flow Models
TripoSG generates high-fidelity 3D meshes from input images via a large-scale rectified flow transformer and hybrid-trained 3D VAE on a custom 2-million-sample dataset, claiming state-of-the-art fidelity and generalization.
-
DVSM: Decoder-only View Synthesis Model Done Right
Decoder-only view synthesis model using KV-cache representation and weight sharing between reconstruction and rendering networks achieves new SOTA on novel view synthesis benchmarks.
-
KFC-W: Generating 3D-Consistent Videos from Unposed Internet Photos
KFC-W is a self-supervised 3D-aware video model trained on videos and multiview internet photos that produces geometrically consistent interpolations between unposed input images without any 3D annotations.
-
AnimateAnyMesh++: A Flexible 4D Foundation Model for High-Fidelity Text-Driven Mesh Animation
AnimateAnyMesh++ animates arbitrary 3D meshes from text using an expanded 300K-identity DyMesh-XL dataset, a power-law topology-aware DyMeshVAE-Flex, and a variable-length rectified-flow generator to produce semantically accurate, temporally coherent animations in seconds.
- Stream3D: Sequential Multi-View 3D Generation via Evidential Memory