SpatialBench evaluates 41 spatial foundation models across 6 paradigms and 5 task suites, finds they are not all-round players, and introduces the DA-Next-5M dataset plus DA-Next baseline model.
hub Mixed citations
$\pi^3$: Permutation-Equivariant Visual Geometry Learning
Mixed citation behavior. Most common role is background (57%).
99
Pith papers citing it
1
external citations · Pith
Background
57% of classified citations
abstract
We introduce $\pi^3$, a feed-forward neural network that offers a novel approach to visual geometry reconstruction, breaking the reliance on a conventional fixed reference view. Previous methods often anchor their reconstructions to a designated viewpoint, an inductive bias that can lead to instability and failures if the reference is suboptimal. In contrast, $\pi^3$ employs a fully permutation-equivariant architecture to predict affine-invariant camera poses and scale-invariant local point maps without any reference frames. This design not only makes our model inherently robust to input ordering, but also leads to higher accuracy and performance. These advantages enable our simple and bias-free approach to achieve state-of-the-art performance on a wide range of tasks, including camera pose estimation, monocular/video depth estimation, and dense point map reconstruction. Code and models are available at https://github.com/yyfz/Pi3.
hub tools
citation-role summary
background 8
method 5
baseline 1
citation-polarity summary
representative citing papers
DrivingDepth achieves SOTA metric depth on nuScenes by residual pixel-wise scale correction on frozen foundation models using sparse LiDAR prompts, preserving geometric consistency.
CasaMaestro predicts metric depth and poses from sparse multi-view panoramas to enable fast house-scale 3D reconstruction.
A new dataset of 220k+ cross-view pairs and a single-stage geometry-aware model GAGeo based on the π³ 3D foundation model outperforms prior methods on object geo-localization with strong generalization and zero-shot ground-to-drone capability.
World Tracing introduces a multi-layer pixel-aligned 3D point representation instantiated via a diffusion transformer (WT-DiT) trained with pixel-space flow matching to jointly reconstruct visible surfaces and generate occluded geometry.
ZipSplat uses multi-view token extraction followed by k-means clustering and attention to decode compact scene tokens into unconstrained 3D Gaussians, achieving SOTA pose-free results with ~6x fewer primitives.
MetricScenes dataset from web photos and stereo imagery, plus a two-stage Poisson depth completion method, allows fine-tuning MoGe-2 to mitigate scale-collapse in metric monocular geometry while preserving benchmark performance.
Geo-Align applies RL with a perceptual reward derived from 3D camera trajectory estimation to improve controllability and fidelity in video generation without paired training data.
Depth2Pose is a new evaluation framework for monocular depth estimators that uses relative camera pose accuracy as a task-driven proxy and introduces the D2P dataset of challenging out-of-distribution scenes.
Mamba-VGGT introduces a Sliding Window Mamba memory module and Zero-Init Spatial Memory Injector to enable persistent long-range geometric reasoning in VGGT for extended video sequences.
A feature-free monocular VINS initialization method that uses feed-forward 3D model point cloud predictions achieves over 90% success rate with under 1.2 seconds of data and performs robustly in degraded environments.
VGGT-Edit proposes a native 3D text-conditioned editing framework using depth-synchronized injection and residual field prediction, plus the DeltaScene dataset, outperforming 2D-lifting methods.
Cross3R performs feed-forward 3D reconstruction and 6-DoF pose estimation from any combination of satellite, UAV, and ground images, outperforming baselines on a new 278K-image tri-view dataset.
CAL2M achieves calibration-free kilometer-level SLAM by using an assistant eye for scale, epipolar-guided intrinsic correction, and anchor propagation for nonlinear sub-map alignment.
A 3D-grounded autoencoder and diffusion transformer allow direct generation of 3D scenes in an implicit latent space using a fixed 1K-token representation for arbitrary views and resolutions.
AnyImageNav uses a semantic-to-geometric cascade with 3D multi-view foundation models to recover precise 6-DoF poses from goal images, achieving 0.27m position error and state-of-the-art success rates on Gibson and HM3D benchmarks.
3D-Fixer performs in-place 3D asset completion from single-view partial point clouds via coarse-to-fine generation with ORFA conditioning, plus a new ARSG-110K dataset, to achieve higher geometric accuracy than MIDI and Gen3DSR while keeping diffusion efficiency.
VGGT-360 delivers geometry-consistent zero-shot panoramic depth by converting panoramas into multi-view 3D reconstructions via VGGT models and three plug-and-play correction modules, then reprojecting the result.
MapAnything is a unified feed-forward transformer that regresses metric 3D scene geometry and cameras from images using a factored representation of depth maps, ray maps, poses, and scale.
FastVGGT achieves 4x speedup on VGGT for 1000-image inputs using training-free token merging tailored to 3D architectures while reducing error accumulation.
A feed-forward feature-Gaussian plus one-step geometry-aware pixel-flow simulator converts large image collections into 20K interactive scenes and 10M+ navigation samples that improve zero-shot Habitat and real-robot performance.
Anchoring Gaussian centers to predicted raymaps and jointly optimizing RGB, raymap, and camera losses with a dual-frequency curriculum suppresses pose drift and improves pose-free 3D reconstruction on long sequences.
TRIG factorizes multi-camera poses into ego-trajectory and static rig geometry, with decoupled supervision and sparse temporal-spatial attention, claiming SOTA metric depth, pose, and 3D reconstruction on five driving benchmarks.
A single pixel-space diffusion model jointly performs 3D scene reconstruction and generation by supervising flow matching on rendered multi-view images, matching SOTA reconstruction and outperforming latent-space generation.
citing papers explorer
-
SpatialBench: Is Your Spatial Foundation Model an All-Round Player?
SpatialBench evaluates 41 spatial foundation models across 6 paradigms and 5 task suites, finds they are not all-round players, and introduces the DA-Next-5M dataset plus DA-Next baseline model.
-
DrivingDepth: Sparse-Prompted Pixel-wise Scale Correction for Driving Depth Estimation
DrivingDepth achieves SOTA metric depth on nuScenes by residual pixel-wise scale correction on frozen foundation models using sparse LiDAR prompts, preserving geometric consistency.
-
CasaMaestro: Multi-View Panoramas for House-Scale 3D Reconstruction
CasaMaestro predicts metric depth and poses from sparse multi-view panoramas to enable fast house-scale 3D reconstruction.
-
Beyond 2D Matching: A Unified Single-Stage Framework for Geometry-Aware Cross-View Object Geo-Localization
A new dataset of 220k+ cross-view pairs and a single-stage geometry-aware model GAGeo based on the π³ 3D foundation model outperforms prior methods on object geo-localization with strong generalization and zero-shot ground-to-drone capability.
-
World Tracing: Generative Pixel-Aligned Geometry Beyond the Visible
World Tracing introduces a multi-layer pixel-aligned 3D point representation instantiated via a diffusion transformer (WT-DiT) trained with pixel-space flow matching to jointly reconstruct visible surfaces and generate occluded geometry.
-
ZipSplat: Fewer Gaussians, Better Splats
ZipSplat uses multi-view token extraction followed by k-means clustering and attention to decode compact scene tokens into unconstrained 3D Gaussians, achieving SOTA pose-free results with ~6x fewer primitives.
-
Honey, I Shrunk the Arc de Triomphe!
MetricScenes dataset from web photos and stereo imagery, plus a two-stage Poisson depth completion method, allows fine-tuning MoGe-2 to mitigate scale-collapse in metric monocular geometry while preserving benchmark performance.
-
Geo-Align: Video Generation Alignment via Metric Geometry Reward
Geo-Align applies RL with a perceptual reward derived from 3D camera trajectory estimation to improve controllability and fidelity in video generation without paired training data.
-
Depth2Pose: A Pose-Based Benchmark for Monocular Depth Estimation without Ground-Truth Depth
Depth2Pose is a new evaluation framework for monocular depth estimators that uses relative camera pose accuracy as a task-driven proxy and introduces the D2P dataset of challenging out-of-distribution scenes.
-
Mamba-VGGT: Persistent Long-Sequence Video Geometry Grounded Transformer via External Sliding Window Mamba Memory
Mamba-VGGT introduces a Sliding Window Mamba memory module and Zero-Init Spatial Memory Injector to enable persistent long-range geometric reasoning in VGGT for extended video sequences.
-
Efficient Feature-Free Initialization for Monocular Visual-Inertial Systems Using a Feed-Forward 3D Model
A feature-free monocular VINS initialization method that uses feed-forward 3D model point cloud predictions achieves over 90% success rate with under 1.2 seconds of data and performs robustly in degraded environments.
-
VGGT-Edit: Feed-forward Native 3D Scene Editing with Residual Field Prediction
VGGT-Edit proposes a native 3D text-conditioned editing framework using depth-synchronized injection and residual field prediction, plus the DeltaScene dataset, outperforming 2D-lifting methods.
-
Seeing Across Skies and Streets: Feedforward 3D Reconstruction from Satellite, Drone, and Ground Images
Cross3R performs feed-forward 3D reconstruction and 6-DoF pose estimation from any combination of satellite, UAV, and ground images, outperforming baselines on a new 278K-image tri-view dataset.
-
Keep It CALM: Toward Calibration-Free Kilometer-Level SLAM with Visual Geometry Foundation Models via an Assistant Eye
CAL2M achieves calibration-free kilometer-level SLAM by using an assistant eye for scale, epipolar-guided intrinsic correction, and anchor propagation for nonlinear sub-map alignment.
-
Any 3D Scene is Worth 1K Tokens: 3D-Grounded Representation for Scene Generation at Scale
A 3D-grounded autoencoder and diffusion transformer allow direct generation of 3D scenes in an implicit latent space using a fixed 1K-token representation for arbitrary views and resolutions.
-
AnyImageNav: Any-View Geometry for Precise Last-Meter Image-Goal Navigation
AnyImageNav uses a semantic-to-geometric cascade with 3D multi-view foundation models to recover precise 6-DoF poses from goal images, achieving 0.27m position error and state-of-the-art success rates on Gibson and HM3D benchmarks.
-
3D-Fixer: Coarse-to-Fine In-place Completion for 3D Scenes from a Single Image
3D-Fixer performs in-place 3D asset completion from single-view partial point clouds via coarse-to-fine generation with ORFA conditioning, plus a new ARSG-110K dataset, to achieve higher geometric accuracy than MIDI and Gen3DSR while keeping diffusion efficiency.
-
VGGT-360: Geometry-Consistent Zero-Shot Panoramic Depth Estimation
VGGT-360 delivers geometry-consistent zero-shot panoramic depth by converting panoramas into multi-view 3D reconstructions via VGGT models and three plug-and-play correction modules, then reprojecting the result.
-
MapAnything: Universal Feed-Forward Metric 3D Reconstruction
MapAnything is a unified feed-forward transformer that regresses metric 3D scene geometry and cameras from images using a factored representation of depth maps, ray maps, poses, and scale.
-
FastVGGT: Training-Free Acceleration of Visual Geometry Transformer
FastVGGT achieves 4x speedup on VGGT for 1000-image inputs using training-free token merging tailored to 3D architectures while reducing error accumulation.
-
Image2Sim: Scaling Embodied Navigation via Generative Neural Simulator
A feed-forward feature-Gaussian plus one-step geometry-aware pixel-flow simulator converts large image collections into 20K interactive scenes and 10M+ navigation samples that improve zero-shot Habitat and real-robot performance.
-
NoDrift3R: Raymap-Guided Coupling for Drift-Robust Unposed Feed-Forward 3D Reconstruction
Anchoring Gaussian centers to predicted raymaps and jointly optimizing RGB, raymap, and camera losses with a dual-frequency curriculum suppresses pose drift and improves pose-free 3D reconstruction on long sequences.
-
TRIG: Trajectory-Rig Decoupled Metric Geometry Learning
TRIG factorizes multi-camera poses into ego-trajectory and static rig geometry, with decoupled supervision and sparse temporal-spatial attention, claiming SOTA metric depth, pose, and 3D reconstruction on five driving benchmarks.
-
PixWorld: Unifying 3D Scene Generation and Reconstruction in Pixel Space
A single pixel-space diffusion model jointly performs 3D scene reconstruction and generation by supervising flow matching on rendered multi-view images, matching SOTA reconstruction and outperforming latent-space generation.
-
The Turning Point of 3D Plant Phenotyping: 3D Foundation Models Enable Minute-to-Second Cross-Crop Reconstruction and Beyond
3D foundation models enable fast cross-crop 3D reconstruction and phenotyping from sparse smartphone images, cutting average time from 6.52 minutes to 1.58 seconds across 26 sequences.
-
VOCA: Visual Odometry with Codec Awareness
VOCA is a causal stereo visual odometry system that achieves state-of-the-art performance on compressed streams by exploiting codec awareness.
-
PointSplat: Compact Gaussian Splatting via Human-Centric Prediction
PointSplat infers compact Gaussian splats directly in 3D space from input point sets via ray casting and Point-Image Transformer to reduce inter-view redundancy and improve novel-view quality for humans.
-
HiReFF: High-Resolution Feedforward Human Reconstruction from Uncalibrated Sparse-View Video
HiReFF presents a feed-forward framework for 2K human video reconstruction from uncalibrated sparse-view videos via scale-synchronized calibration, Gaussian masking, and high-resolution side-tuning.
-
TryOnCrafter: Unleashing Camera Trajectories for Realistic Video Virtual Try-on via a Renderable 4D Try-on Proxy
A 4D try-on proxy (3DGS avatar + SMPL-X + background points) anchors a DiT so virtual try-on videos can follow arbitrary camera trajectories with consistent garments and scene structure.
-
OrthoTrack: Continuous 6-DoF UAV Trajectory Estimation Anchored in Public Orthophotos
OrthoTrack is a training-free system for continuous metric 6-DoF UAV pose estimation anchored in public orthophotos and surface models, with a new MovingDrone benchmark dataset.
-
G$^3$VLA: Geometric inductive bias for Vision-Language-Action Models
G³VLA injects calibrated camera geometry into VLA visual tokens via intrinsic-conditioned ray embeddings, PRoPE, and bidirectional cross-view fusion, producing consistent gains on LIBERO, RoboCasa24, RoboTwin2.0, and real-robot tasks when added to π₀.
-
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.
-
MV-Actor: Aligning Multi-View Semantics and Spatial Awareness for Bimanual Manipulation
MV-Actor proposes a multi-view framework using semantic interaction, semantic-spatial token interaction, and guided depth repair to reach 87.8% success on the PerAct2 bimanual benchmark and outperform RGB/RGB-D baselines in real-world tests.
-
Anchor3R: Streaming 3D Reconstruction with Transient Anchors for Long-Horizon Visual Mapping
Anchor3R reframes feed-forward 3D reconstruction as current-centric local measurement prediction, using loop-closure and motion averaging to produce coherent global maps from visual streams.
-
Modeling Depth Ambiguity: A Mixture-Density Representation for Flying-Point-Free Depth Estimation
MDA represents per-pixel depth as a mixture of distributions so that boundary pixels can align hypotheses with distinct surfaces instead of averaging into empty space.
-
DeblurNVS: Geometric Latent Diffusion for Novel View Synthesis from Sparse Motion-Blurred Images
DeblurNVS restores geometric representations via latent diffusion to enable high-fidelity novel view synthesis directly from sparse motion-blurred inputs.
-
RayDer: Scalable Self-Supervised Novel View Synthesis from Real-World Video
RayDer is a unified transformer backbone for self-supervised static-scene novel view synthesis that absorbs dynamic content as a nuisance factor and shows power-law scaling with data and compute while matching supervised methods in zero-shot settings.
-
D\'ej\`a View: Looping Transformers for Multi-View 3D Reconstruction
DéjàView applies a single transformer block recurrently for K refinement steps, matching or exceeding larger feed-forward models on five multi-view 3D benchmarks with fewer parameters and comparable compute.
-
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.
-
Sensor2Sensor: Cross-Embodiment Sensor Conversion for Autonomous Driving
Sensor2Sensor uses 4D Gaussian Splatting to create synthetic training pairs and a diffusion model to convert monocular dashcam videos into high-fidelity multi-modal AV sensor data.
-
UniT: Unified Geometry Learning with Group Autoregressive Transformer
UniT unifies online and offline 3D geometry perception via a Group Autoregressive Transformer that processes observation groups with anchor-free point map prediction and a scale-adaptive loss.
-
How You Move Tells What You'll Do: Trajectory-Conditioned Egocentric Prediction
TrajPilot predicts candidate future trajectories from egocentric context and uses them to condition action prediction in an embedding space, outperforming VLM and planner baselines on Ego-Exo4D, Ego4D, and other datasets with gains increasing at longer horizons.
-
Cross-View Splatter: Feed-Forward View Synthesis with Georeferenced Images
A feed-forward model aligns ground and satellite features to predict Gaussian splats for improved novel-view synthesis on georeferenced outdoor scenes.
-
GeoFlow: Enforcing Implicit Geometric Consistency in Video Generation
GeoFlow adds a geometry-consistency reward based on rigid camera flow and object appearance preservation, integrated via reinforcement fine-tuning to improve geometric coherence in video generation.
-
LongDPM: Overlap-Aware 4D Reconstruction from Long Monocular Videos
LongDPM introduces an overlap-aware chunk-based framework that registers and fuses local dynamic reconstructions to achieve coherent long-range 4D geometry and tracking from monocular video.
-
EndoGSim: Physics-Aware 4D Dynamic Endoscopic Scene Simulations via MLLM-Guided Gaussian Splatting
EndoGSim integrates MLLM-guided material initialization with 4D Gaussian Splatting and differentiable Material Point Method to achieve physics-aware 4D reconstruction and simulation of endoscopic scenes.
-
Unlocking Dense Metric Depth Estimation in VLMs
DepthVLM converts a standard VLM into a dense metric depth predictor by attaching a lightweight head and training under unified vision-text supervision, outperforming prior VLMs and some pure vision models on a new indoor-outdoor benchmark.
-
Warp-as-History: Generalizable Camera-Controlled Video Generation from One Training Video
Warp-as-History enables zero-shot camera trajectory following in frozen video models by supplying camera-warped pseudo-history, with single-video LoRA fine-tuning improving generalization to unseen videos.
-
Spark3R: Asymmetric Token Reduction Makes Fast Feed-Forward 3D Reconstruction
Spark3R achieves up to 28x speedup on 1000-frame 3D reconstruction inputs by asymmetrically reducing query and key-value tokens in Vision Transformers while keeping competitive quality.
-
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.