InstantHMR recovers 3D human meshes in real time on phones
A Galaxy S23 NPU produces a 6,890-vertex mesh in 5.5 ms, with 40.0 mm PA-MPJPE on 3DPW.
Stories per day · 30 days
3 RSS
Relatedmulti-view2procedural-generation2video-generation2agents1blender1cinemabench1cinematography1computer-vision1diffusion1filmflow1game-assets1hair-generation1human-mesh-recovery1infinidream1mobile-inference1open-source1planning1robotics1
27 stories, newest first · one card per story
Stories tagged #3d
HairDeck generates animatable stylized hair cards from one image
On CHARM's test set it lowers Chamfer distance by 44% and card-orientation error by more than half, while running 33 times faster.
OmniShowcase generates subject-consistent video from multi-view images
Training covers 12 camera motions and 88 trajectories, including full 360-degree orbits, with surface-normal conditioning from reference views.
FilmFlow plans 3D camera paths from text or reference video
CinemaBench covers 150 Blender scenes, 750 text-directed tasks, and 60 reference videos.
WorldStream streams open-ended 3D worlds with persistent editable state
Hierarchical planners compile terrain ahead of the player while one world record restores prior visits and edits.
Video-Point Model forecasts multi-view video and 3D point trajectories
Point supervision on VPM-Data and VP-WorldBench improves video fidelity, 3D motion, and robot manipulation success.
InfiniDream generates editable photorealistic 3D worlds with VLMs
Evaluations on 100 object and 20 scene tasks report gains on four perceptual metrics.
OmniCam generates camera trajectories from a panorama and text
On OmniCaT, reported trajectory error falls 28–47% and collisions 65.8% versus retrained GenDoP.
MAMHOI: Factorizing Scene-Aware Human-Object Interaction through Affordances
MAMHOI factorizes scene-aware human-object interaction through affordances, reducing penetration without paired scene data.
SpaceCast-Bench: Evaluating Predictive Spatial Reasoning in Vision-Language Models
SpaceCast-Bench tests predictive spatial reasoning; the best of 21 models scores 58.0% versus 87.2% for humans.
MAMHOI: Factorizing Scene-Aware Human-Object Interaction through Affordances
MAMHOI factorizes scene-aware human-object interaction generation through motion affordances without paired scene data.
SpaceCast-Bench: Evaluating Predictive Spatial Reasoning in Vision-Language Models
SpaceCast-Bench finds VLMs weak at predictive spatial reasoning; the best model scores 58 percent versus 87.2 percent for humans.
Distilling Routed 3D Privilege for Spatial Reasoning in Vision-Language Models
GPD distills routed 3D geometry into an RGB-only VLM, lifting a 4B model to 57.1 on VSI-Bench.
Pumpire: Unified Benchmark for Metric Distance Estimation
Pumpire benchmarks point-pair metric distance accuracy for image and video 3D foundation models.
DepthWorld: 3D World Model for Robot Manipulation
DepthWorld jointly predicts multi-view RGB and metric depth for robot manipulation from calibrated 3D data.
CADFather: Autonomous CAD Reconstruction through Coordinated Tool Use
CADFather coordinates tools and a vision-language agent to reconstruct editable CAD programs from 3D meshes.
NAMVIS: Next-Scale Autoregressive Multi-View Image Synthesis
NAMVIS generates consistent multi-view images with geometry-conditioned autoregression, beating diffusion baselines and running over 3x faster.
Magic-W0: A Structured World-Action Foundation Model for Physical Intelligence
Magic-W0 jointly models structured 3D world transitions and robot actions, leading compared world-action models.
PointWAM: 3D World Action Modeling for Dexterous Robotic Manipulation
PointWAM forecasts 3D scene and hand trajectories, lifting dexterous robot success on DexJoCo tasks.
Generative Cinematographer: Composing Camera and Object Motion in 3D
GenCine lets artists jointly control 3D camera and object motion in video via guidance maps on a Wan model.
4Director: Controlling Video World Models with Rigid 3D Geometry
4Director controls video world models using rigid meshes and per-frame transforms for consistent camera and object motion.
World Observer: Joint Actor-Observer Generation for Persistent World Modeling
World Observer jointly generates an actor view and panoramic observers so video world models preserve out-of-view object dynamics.
WorldAuditBench: Interactive 3D World Auditing with Multimodal Agents
WorldAuditBench tests multimodal agents at finding 3D-world anomalies; frontier models reach 6.6–42.3% versus humans at 83.4%.
WorldAuditBench: Interactive 3D World Auditing with Multimodal Agents
WorldAuditBench shows multimodal agents trail humans when auditing anomalies in interactive 3D worlds.