3DV 2025

150 papers
The 2025 edition of the 3DV conference featured 190 papers focusing on advancements in 3D scene decomposition and real-time spatial perception. Key research directions included the development of decoupled architectures for spatial awareness and the application of geometric priors for structural hallucination. Technical discussions also emphasized the sensitivity of 3D reconstruction processes to initial segmentation stages. Methodological trends highlighted the integration of diffusion models for 3D object generation and human modeling, as evidenced by research into joint-aware latent diffusion and co-speech gesture generation. The proceedings showcased progress in zero-shot instance segmentation via 3D prompt localization and efficient view synthesis using 3D Gaussian Splatting. Other contributions addressed large-scale structure from motion and 3D registration techniques. Prolific authors contributing to the year's program included researchers such as Andrea Vedaldi, Christian Theobalt, and Luc Van Gool.

Top Research Directions

Research directions most represented by 3DV 2025 papers.

Physics Constrained Generative Human Motion Synthesis 4
Resolving Shape Radiance Ambiguity in Sparse View Reconstruction 4
Overcoming the Fixed Topology Constraint in 3D Modeling 3
Editing Neural Implicit Representations 2
Pruning Redundant Primitives in 3D Gaussian Splatting 2
Real-Time Autoregressive Generation for Interactive Human Motion 2
Generative Simulation for Scaling Embodied AI 2
Discrete Tokenization for Human Motion Modeling 2
Mitigating Multi View Inconsistency in Generative 3D Modeling 2
Overcoming the Static Scene Assumption 2
Overcoming the Static Scene Assumption in Computer Vision 2
Beyond the Lambertian Assumption for Non-Diffuse Surfaces 2
Accelerating Neural Scene Representations for Real Time Rendering 2
Resolving Geometric Ambiguity in Sparse View Reconstruction 2
Material Lighting Disentanglement for 3D Reconstruction 2

Top Authors

AuthorPapers
Andrea Vedaldi4
Christian Theobalt4
Gerard Pons-Moll4
Shenlong Wang4
Vladislav Golyanik4
Angel X Chang3
Bernhard Thomaszewski3
Chuhang Zou3
Deva Ramanan3
Dimitris Samaras3
Eddy Ilg3
Federico Tombari3
Junlin Han3
Justus Thies3
Luc Van Gool3
Michael J. Black3
Otmar Hilliges3
Ping Tan3
Artur Grigorev2
Bardienus Pieter Duisterhof2

Impactful Papers

Papers from 3DV 2025 that shaped research clusters.

Foundational

Influential

Papers

Title
Para-Lane: Multi-Lane Dataset Registering Parallel Scans for Benchmarking Novel View Synthesis
Particle Rendering: Implicitly Aggregating Incident and Outgoing Light Fields for Novel View Synthesis
PIR: Photometric Inverse Rendering with Shading Cues Modeling and Surface Reflectance Regularization
Plenoptic PNG: Real-Time Neural Radiance Fields in 150 KB
Purposer: Putting Human Motion Generation in Context
Pushing the Limits of LiDAR: Accurate Performance Analysis of Indoor 3D LiDARs
RadSplat: Radiance Field-Informed Gaussian Splatting for Robust Real- Time Rendering with 900+ FPS
RadSplat: Radiance Field-Informed Gaussian Splatting for Robust Real-Time Rendering with 900+ FPS
RealmDreamer: Text-Driven 3D Scene Generation with Inpainting and Depth Diffusion
Reason3D: Searching and Reasoning 3D Segmentation via Large Language Model
Reflecting Reality: Enabling Diffusion Models to Produce Faithful Mirror Reflections
Rig3DGS: Creating Controllable Portraits From Casual Monocular Videos
Rigid Body Adversarial Attacks
RISE-SDF: A Relightable Information-Shared Signed Distance Field for Glossy Object Inverse Rendering
Robustifying Point Cloud Networks by Refocusing
Robust Spectral Translation Synchronization
SAIL-Recon: Large SfM by Augmenting Scene Regression with Localization
SAMPro3D: Locating SAM Prompts in 3D for Zero-Shot Instance Segmentation
SAT-SKYLINES: 3D Building Generation from Satellite Imagery and Coarse Geometric Priors
SceneMotifCoder: Example-driven Visual Program Learning for Generating 3D Object Arrangements
ShadowSG: Spherical Gaussian Illumination from Shadows
SH-SAS: An Implicit Neural Representation for Complex Spherical-Harmonic Scattering Fields for 3D Synthetic Aperture Sonar
SMORE: Simultaneous Map and Object REconstruction
Snap-it, Tap-it, Splat-it: Tactile-Informed 3D Gaussian Splatting for Reconstructing Challenging Surfaces
SocialGen: Modeling Multi-Human Social Interaction with Language Models
SPAFormer: Sequential 3D Part Assembly with Transformers
SparseGS: Sparse View Synthesis using 3D Gaussian Splatting
Spatial Cognition from Egocentric Video: Out of Sight, Not Out of Mind
SphereFusion: Efficient Panorama Depth Estimation via Gated Fusion
Spurfies: Sparse-View Surface Reconstruction using Local Geometry Priors
SSRFlow: Semantic-Aware Fusion with Spatial Temporal Re-Embedding for Real-World Scene Flow
SurfR: Surface Reconstruction with Multi-scale Attention
Synthesizing Consistent Novel Views Via 3D Epipolar Attention Without Re-Training
TEDRA: Text-Based Editing of Dynamic and Photoreal Actors
TexAvatars: Hybrid Texel-3D Representations for Stable Rigging of Photorealistic Gaussian Head Avatars
Towards Foundation Models for 3D Vision: How Close Are We?
TTT-KD: Test-Time Training for 3D Semantic Segmentation Through Knowledge Distillation From Foundation Models
U–ARE–ME: Uncertainty-Aware Rotation Estimation in Manhattan Environments
UniMotion: Unifying 3D Human Motion Synthesis and Understanding
UNIT: Unsupervised Online Instance Segmentation Through Time
UrbanIR: Large-Scale Urban Scene Inverse Rendering from a Single Video
ViSkin: Physics-based Simulation of Virtual Skin on Personalized Avatars
ViSTA-SLAM: Visual SLAM with Symmetric Two-view Association
VividTalk: One-Shot Audio-Driven Talking Head Generation Based on 3D Hybrid Prior
Vocabulary-Free 3D Instance Segmentation with Vision-Language Assistant
VXP: Voxel-Cross-Pixel Large-Scale Camera-LiDAR Place Recognition
WaterSplatting: Fast Underwater 3D Scene Reconstruction using Gaussian Splatting
XLD: A Cross-Lane Dataset for Benchmarking Novel Driving View Synthesis
ZeroPS: High-quality Cross-modal Knowledge Transfer for Zero-Shot 3D Part Segmentation
αSurf: Implicit Surface Reconstruction for Semi-Transparent and Thin Objects with Decoupled Geometry and Opacity

« prev Page 3/3