CVPR 2025
自动驾驶 · 具身智能论文归档

Enduring, Efficient and Robust Trajectory Prediction Attack in Autonomous Driving via Optimization-Driven Multi-Frame Perturbation Framework

DiffusionDrive: Truncated Diffusion Model for End-to-End Autonomous Driving

MPDrive: Improving Spatial Understanding with Marker-Based Prompt Learning for Autonomous Driving

DriveGPT4-V2: Harnessing Large Language Model Capabilities for Enhanced Closed-Loop Autonomous Driving

Driving by the Rules: A Benchmark for Integrating Traffic Sign Regulations into Vectorized HD Map

SimLingo: Vision-Only Closed-Loop Autonomous Driving with Language-Action Alignment

BEVDiffuser: Plug-and-Play Diffusion Model for BEV Denoising with Ground-Truth Guidance

Towards Explicit Geometry-Reflectance Collaboration for Generalized LiDAR Segmentation in Adverse Weather

FoundationStereo: Zero-Shot Stereo Matching

Continuous 3D Perception Model with Persistent State

TacoDepth: Towards Efficient Radar-Camera Depth Estimation with One-stage Fusion

Keep the Balance: A Parameter-Efficient Symmetrical Framework for RGB+X Semantic Segmentation

DORNet: A Degradation Oriented and Regularized Network for Blind Depth Super-Resolution

Learned Binocular-Encoding Optics for RGBD Imaging Using Joint Stereo and Focus Cues

Opportunistic Single-Photon Time of Flight

Zero-Shot Monocular Scene Flow Estimation in the Wild

Effective SAM Combination for Open-Vocabulary Semantic Segmentation

Towards RAW Object Detection in Diverse Conditions

Scene-Centric Unsupervised Panoptic Segmentation

MITracker: Multi-View Integration for Visual Object Tracking

CrossOver: 3D Scene Cross-Modal Alignment

Flash3D: Super-scaling Point Transformers through Joint Hardware-Geometry Locality

EventPSR: Surface Normal and Reflectance Estimation from Photometric Stereo Using an Event Camera

A Unified Approach to Interpreting Self-supervised Pre-training Methods for 3D Point Clouds via Interactions

SAMWISE: Infusing Wisdom in SAM2 for Text-Driven Video Segmentation

One-shot 3D Object Canonicalization based on Geometric and Semantic Consistency

WISH: Weakly Supervised Instance Segmentation using Heterogeneous Labels

Change3D: Revisiting Change Detection and Captioning from A Video Modeling Perspective

Learning 4D Panoptic Scene Graph Generation from Rich 2D Visual Scene

v-CLR: View-Consistent Learning for Open-World Instance Segmentation

LLMDet: Learning Strong Open-Vocabulary Object Detectors under the Supervision of Large Language Models

Task-driven Image Fusion with Learnable Fusion Loss

Video Depth Anything: Consistent Depth Estimation for Super-Long Videos

Cross-View Completion Models are Zero-shot Correspondence Estimators

MonSter: Marry Monodepth to Stereo Unleashes Power

Prior-free 3D Object Tracking

CH3Depth: Efficient and Flexible Depth Foundation Model with Flow Matching

MambaVLT: Time-Evolving Multimodal State Space Model for Vision-Language Tracking

ETAP: Event-based Tracking of Any Point

Graph Neural Network Combining Event Stream and Periodic Aggregation for Low-Latency Event-based Vision

LP-Diff: Towards Improved Restoration of Real-World Degraded License Plate

Text-guided Sparse Voxel Pruning for Efficient 3D Visual Grounding

All-directional Disparity Estimation for Real-world QPD Images

Detection-Friendly Nonuniformity Correction: A Union Framework for Infrared UAV Target Detection

Mamba as a Bridge: Where Vision Foundation Models Meet Vision Language Models for Domain-Generalized Semantic Segmentation

Cubify Anything: Scaling Indoor 3D Object Detection

Ev-3DOD: Pushing the Temporal Boundaries of 3D Object Detection with Event Cameras

Seurat: From Moving Points to Depth

SP3D: Boosting Sparsely-Supervised 3D Object Detection via Accurate Cross-Modal Semantic Prompts

Learning Class Prototypes for Unified Sparse-Supervised 3D Object Detection

Event Fields: Capturing Light Fields at High Speed, Resolution, and Dynamic Range

CRISP: Object Pose and Shape Estimation with Test-Time Adaptation

Sonata: Self-Supervised Learning of Reliable Point Representations

Motion Prompting: Controlling Video Generation with Motion Trajectories

Go-with-the-Flow: Motion-Controllable Video Diffusion Models Using Real-Time Warped Noise

Temporally Consistent Object-Centric Learning by Contrasting Slots

Reanimating Images using Neural Representations of Dynamic Stimuli

Navigation World Models

GEN3C: 3D-Informed World-Consistent Video Generation with Precise Camera Control

FIction: 4D Future Interaction Prediction from Video

World-consistent Video Diffusion with Explicit 3D Modeling

EgoLM: Multi-Modal Language Model of Egocentric Motions

From Multimodal LLMs to Generalist Embodied Agents: Methods and Lessons

PDFactor: Learning Tri-Perspective View Policy Diffusion Field for Multi-Task Robotic Manipulation

RoboSpatial: Teaching Spatial Understanding to 2D and 3D Vision-Language Models for Robotics

InteractAnything: Zero-shot Human Object Interaction Synthesis via LLM Feedback and Object Affordance Parsing

SkillMimic: Learning Basketball Interaction Skills from Demonstrations

CAP-Net: A Unified Network for 6D Pose and Size Estimation of Categorical Articulated Parts from a Single RGB-D Image

ManiVideo: Generating Hand-Object Manipulation Video with Dexterous and Generalizable Grasping

Generating 6DoF Object Manipulation Trajectories from Action Description in Egocentric Vision

OmniManip: Towards General Robotic Manipulation via Object-Centric Interaction Primitives as Spatial Constraints

How Do I Do That? Synthesizing 3D Hand Motion and Contacts for Everyday Interactions

End-to-End HOI Reconstruction Transformer with Graph-based Encoding

Shape Abstraction via Marching Differentiable Support Functions

DexGrasp Anything: Towards Universal Robotic Dexterous Grasping with Physics Awareness

Self-Supervised Cross-View Correspondence with Predictive Cycle Consistency

RoboPEPP: Vision-Based Robot Pose and Joint Angle Estimation through Embedding Predictive Pre-Training

CheckManual: A New Challenge and Benchmark for Manual-based Appliance Manipulation

TokenHSI: Unified Synthesis of Physical Human-Scene Interactions through Task Tokenization

GROVE: A Generalized Reward for Learning Open-Vocabulary Physical Skill

Estimating Body and Hand Motion in an Ego-sensed World

Lifting Motion to the 3D World via 2D Diffusion

EnergyMoGen: Compositional Human Motion Generation with Energy-Based Diffusion Model in Latent Space

ClimbingCap: Multi-Modal Dataset and Method for Rock Climbing in World Coordinate

UMotion: Uncertainty-driven Human Motion Estimation from Inertial and Ultra-wideband Units

H-MoRe: Learning Human-centric Motion Representation for Action Analysis

ALIEN: Implicit Neural Representations for Human Motion Prediction under Arbitrary Latency

HSI-GPT: A General-Purpose Large Scene-Motion-Language Model for Human Scene Interaction

InterMimic: Towards Universal Whole-Body Control for Physics-Based Human-Object Interactions

MotionPRO: Exploring the Role of Pressure in Human MoCap and Beyond

FRAME: Floor-aligned Representation for Avatar Motion from Egocentric Video

Thinking in Space: How Multimodal Large Language Models See, Remember, and Recall Spaces

Inst3D-LMM: Instance-Aware 3D Scene Understanding with Multi-modal Instruction Tuning

Functionality Understanding and Segmentation in 3D Scenes

Reasoning in Visual Navigation of End-to-end Trained Agents: A Dynamical Systems Approach

DIV-FF: Dynamic Image-Video Feature Fields For Environment Understanding in Egocentric Videos

Open-Vocabulary Functional 3D Scene Graphs for Real-World Indoor Spaces

Descriptor-In-Pixel : Point-Feature Tracking For Pixel Processor Arrays

VGGT: Visual Geometry Grounded Transformer

MegaSaM: Accurate, Fast and Robust Structure and Motion from Casual Dynamic Videos

Convex Relaxation for Robust Vanishing Point Estimation in Manhattan World

Camera Resection from Known Line Pencils and a Radially Distorted Scanline

Gromov–Wasserstein Problem with Cyclic Symmetry

SSHNet: Unsupervised Cross-modal Homography Estimation via Problem Reformulation and Split Optimization

MUSt3R: Multi-view Network for Stereo 3D Reconstruction

Structure-Aware Correspondence Learning for Relative Pose Estimation

MASt3R-SLAM: Real-Time Dense SLAM with 3D Reconstruction Priors

Can Generative Video Models Help Pose Estimation?

SLAM3R: Real-Time Dense Scene Reconstruction from Monocular RGB Videos

Unlocking Generalization Power in LiDAR Point Cloud Registration

Learning to Filter Outlier Edges in Global SfM

Doppelgangers++: Improved Visual Disambiguation with Geometric 3D Features

Full-DoF Egomotion Estimation for Event Cameras Using Geometric Solvers

Structure-from-Motion with a Non-Parametric Camera Model

BADGR: Bundle Adjustment Diffusion Conditioned by Gradients for Wide-Baseline Floor Plan Reconstruction

MATCHA: Towards Matching Anything

Order-One Rolling Shutter Cameras

HaWoR: World-Space Hand Motion Reconstruction from Egocentric Videos

Relative Pose Estimation through Affine Corrections of Monocular Depth Priors

Align3R: Aligned Monocular Depth Estimation for Dynamic Videos

Light3R-SfM: Towards Feed-forward Structure-from-Motion

IncEventGS: Pose-Free Gaussian Splatting from a Single Event Camera

ForestLPR: LiDAR Place Recognition in Forests Attentioning Multiple BEV Density Images

Implicit Correspondence Learning for Image-to-Point Cloud Registration

EBS-EKF: Accurate and High Frequency Event-based Star Tracking

Closed-Loop Supervised Fine-Tuning of Tokenized Traffic Models

ImagineFSL: Self-Supervised Pretraining Matters on Imagined Base Set for VLM-based Few-shot Learning

COUNTS: Benchmarking Object Detectors and Multimodal Large Language Models under Distribution Shifts

The Scene Language: Representing Scenes with Programs, Words, and Embeddings

Towards Autonomous Micromobility through Scalable Urban Simulation

HOT3D: Hand and Object Tracking in 3D from Egocentric Multi-View Videos

HumanRig: Learning Automatic Rigging for Humanoid Character in a Large Scale Dataset

RoboTwin: Dual-Arm Robot Benchmark with Generative Digital Twins

WonderWorld: Interactive 3D Scene Generation from a Single Image

HELVIPAD: A Real-World Dataset for Omnidirectional Stereo Depth Estimation

EgoPressure: A Dataset for Hand Pressure and Pose Estimation in Egocentric Vision

CASAGPT: Cuboid Arrangement and Scene Assembly for Interior Design

GigaHands: A Massive Annotated Dataset of Bimanual Hand Activities

Digital Twin Catalog: A Large-Scale Photorealistic 3D Object Digital Twin Dataset

OPTICAL: Leveraging Optimal Transport for Contribution Allocation in Dataset Distillation

HuPerFlow: A Comprehensive Benchmark for Human vs. Machine Motion Estimation Comparison

Alias-Free Latent Diffusion Models: Improving Fractional Shift Equivariance of Diffusion Latent Space

RandAR: Decoder-only Autoregressive Visual Generation in Random Orders

CAP4D: Creating Animatable 4D Portrait Avatars with Morphable Multi-View Diffusion Models

MoGe: Unlocking Accurate Monocular Geometry Estimation for Open-Domain Images with Optimal Training Supervision

Multi-view Reconstruction via SfM-guided Monocular Depth Estimation

MV-DUSt3R+: Single-Stage Scene Reconstruction from Sparse Views In 2 Seconds

Reconstructing Humans with a Biomechanically Accurate Skeleton

MEGA: Masked Generative Autoencoder for Human Mesh Recovery

CraftsMan3D: High-fidelity Mesh Generation with 3D Native Diffusion and Interactive Geometry Refiner

Stereo4D: Learning How Things Move in 3D from Internet Stereo Videos

Neural Inverse Rendering from Propagating Light

Time of the Flight of the Gaussians: Optimizing Depth Indirectly in Dynamic Radiance Fields

3D Student Splatting and Scooping

DIFIX3D+: Improving 3D Reconstructions with Single-Step Diffusion Models

3DGUT: Enabling Distorted Cameras and Secondary Rays in Gaussian Splatting

FluidNexus: 3D Fluid Reconstruction and Prediction from a Single Video

Diffusion Renderer: Neural Inverse and Forward Rendering with Video Diffusion Models

Tuning the Frequencies: Robust Training for Sinusoidal Neural Networks

You See it, You Got it: Learning 3D Creation on Pose-Free Videos at Scale

FirePlace: Geometric Refinements of LLM Common Sense Reasoning for 3D Object Placement

3D Convex Splatting: Radiance Field Rendering with 3D Smooth Convexes

ArcPro: Architectural Programs for Structured 3D Abstraction of Sparse Points

GenVDM: Generating Vector Displacement Maps From a Single Image

PlanarSplatting: Accurate Planar Surface Reconstruction in 3 Minutes

Instant Gaussian Stream: Fast and Generalizable Streaming of Dynamic Scene Reconstruction via Gaussian Splatting

MAtCha Gaussians: Atlas of Charts for High-Quality Geometry and Photorealism From Sparse Views

CADDreamer: CAD Object Generation from Single-view Images

Generative Multiview Relighting for 3D Reconstruction under Extreme Illumination Variation

PGC: Physics-Based Gaussian Cloth from a Single Pose

PartGen: Part-level 3D Generation and Reconstruction with Multi-view Diffusion Models

DualPM: Dual Posed-Canonical Point Maps for 3D Shape and Pose Reconstruction

Blurred LiDAR for Sharper 3D: Robust Handheld 3D Scanning with Diffuse LiDAR and RGB

Improving Gaussian Splatting with Localized Points Management

Towards In-the-wild 3D Plane Reconstruction from a Single Image

Real-time Free-view Human Rendering from Sparse-view RGB Videos using Double Unprojected Textures

MetricGrids: Arbitrary Nonlinear Approximation with Elementary Metric Grids based Implicit Neural Representation

High-Fidelity Lightweight Mesh Reconstruction from Point Clouds

Material Anything: Generating Materials for Any 3D Object via Diffusion

Locally Orderless Images for Optimization in Differentiable Rendering

Structure from Collision

Uni4D: Unifying Visual Foundation Models for 4D Modeling from a Single Video

Volumetrically Consistent 3D Gaussian Rasterization

Taming Video Diffusion Prior with Scene-Grounding Guidance for 3D Gaussian Splatting from Sparse Inputs

Unified Reconstruction of Static and Dynamic Scenes from Events

FreeCloth: Free-form Generation Enhances Challenging Clothed Human Modeling

Make-It-Animatable: An Efficient Framework for Authoring Animation-Ready 3D Characters

GaussianUDF: Inferring Unsigned Distance Functions through 3D Gaussian Splatting

SpecTRe-GS: Modeling Highly Specular Surfaces with Reflected Nearby Objects by Tracing Rays in 3D Gaussian Splatting

OmniSplat: Taming Feed-Forward 3D Gaussian Splatting for Omnidirectional Images with Editable Capabilities

Symmetry Strikes Back: From Single-Image Symmetry Detection to 3D Generation

DroneSplat: 3D Gaussian Splatting for Robust 3D Reconstruction from In-the-Wild Drone Imagery

Reconstructing People, Places, and Cameras

VideoScene: Distilling Video Diffusion Model to Generate 3D Scenes in One Step

NeRFPrior: Learning Neural Radiance Field as a Prior for Indoor Scene Reconstruction

3DTopia-XL: Scaling High-quality 3D Asset Generation via Primitive Diffusion

Creating Your Editable 3D Photorealistic Avatar with Tetrahedron-constrained Gaussian Splatting

High-fidelity 3D Object Generation from Single Image with RGBN-Volume Gaussian Reconstruction Model

Dyn-HaMR: Recovering 4D Interacting Hand Motion from a Dynamic Camera

MAR-3D: Progressive Masked Auto-regressor for High-Resolution 3D Generation

Generative Densification: Learning to Densify Gaussians for High-Fidelity Generalizable 3D Reconstruction

HotSpot: Signed Distance Function Optimization with an Asymptotically Sufficient Condition

BWFormer: Building Wireframe Reconstruction from Airborne LiDAR Point Cloud with Transformer

Structured 3D Latents for Scalable and Versatile 3D Generation

Hardware-Rasterized Ray-Based Gaussian Splatting

ARM: Appearance Reconstruction Model for Relightable 3D Generation

Dr. Splat: Directly Referring 3D Gaussian Splatting via Direct Language Embedding Registration

Real-time High-fidelity Gaussian Human Avatars with Position-based Interpolation of Spatially Distributed MLPs

MoSca: Dynamic Gaussian Fusion from Casual Videos via 4D Motion Scaffolds

NexusGS: Sparse View Synthesis with Epipolar Depth Priors in 3D Gaussian Splatting

USP-Gaussian: Unifying Spike-based Image Reconstruction, Pose Correction and Gaussian Splatting

DashGaussian: Optimizing 3D Gaussian Splatting in 200 Seconds

SplatFlow: Self-Supervised Dynamic Gaussian Splatting in Neural Motion Flow Field for Autonomous Driving

Glossy Object Reconstruction with Cost-effective Polarized Acquisition

Matrix3D: Large Photogrammetry Model All-in-One

DepthCrafter: Generating Consistent Long Depth Sequences for Open-world Videos

FRESA: Feedforward Reconstruction of Personalized Skinned Avatars from Few Images

SKDream: Controllable Multi-view and 3D Generation with Arbitrary Skeletons

CleanDIFT: Diffusion Features without Noise

OverLoCK: An Overview-first-Look-Closely-next ConvNet with Context-Mixing Dynamic Kernels

Do We Always Need the Simplicity Bias? Looking for Optimal Inductive Biases in the Wild

DiffFNO: Diffusion Fourier Neural Operator

Molmo and PixMo: Open Weights and Open Data for State-of-the-Art Vision-Language Models

Rethinking Spiking Self-Attention Mechanism: Implementing α-XNOR Similarity Calculation in Spiking Transformers

Efficient Test-time Adaptive Object Detection via Sensitivity-Guided Pruning

Viewpoint Rosetta Stone: Unlocking Unpaired Ego-Exo Videos for View-invariant Representation Learning

UniAP: Unifying Inter- and Intra-Layer Automatic Parallelism by Mixed Integer Quadratic Programming

Enhancing Diversity for Data-free Quantization

LoRASculpt: Sculpting LoRA for Harmonizing General and Specialized Knowledge in Multimodal Large Language Models

Generative Multimodal Pretraining with Discrete Diffusion Timestep Tokens

Context-Aware Multimodal Pretraining

Balanced Rate-Distortion Optimization in Learned Image Compression

ICP: Immediate Compensation Pruning for Mid-to-high Sparsity

MASH-VLM: Mitigating Action-Scene Hallucination in Video-LLMs through Disentangled Spatial-Temporal Representations

Multimodal Autoregressive Pre-training of Large Vision Encoders

Tracktention: Leveraging Point Tracking to Attend Videos Faster and Better

Breaking the Memory Barrier of Contrastive Loss via Tile-Based Strategy

Coeff-Tuning: A Graph Filter Subspace View for Tuning Attention-Based Large Models

Olympus: A Universal Task Router for Computer Vision Tasks

SoMA: Singular Value Decomposed Minor Components Adaptation for Domain Generalizable Representation Learning

SURGEON: Memory-Adaptive Fully Test-Time Adaptation via Dynamic Activation Sparsity

Good, Cheap, and Fast: Overfitted Image Compression with Wasserstein Distortion

Multirate Neural Image Compression with Adaptive Lattice Vector Quantization

CARE Transformer: Mobile-Friendly Linear Visual Transformer via Decoupled Dual Interaction

InPO: Inversion Preference Optimization with Reparametrized DDIM for Efficient Diffusion Model Alignment

SnapGen: Taming High-Resolution Text-to-Image Models for Mobile Devices with Efficient Architectures and Training

Your ViT is Secretly an Image Segmentation Model

Your Large Vision-Language Model Only Needs A Few Attention Heads For Visual Grounding

Meta-Learning Hyperparameters for Parameter Efficient Fine-Tuning

CASP: Compression of Large Multimodal Models Based on Attention Sparsity

Efficient Motion-Aware Video MLLM

FlexiDiT: Your Diffusion Transformer Can Easily Generate High-Quality Samples with Less Compute

TinyFusion: Diffusion Transformers Learned Shallow

FIMA-Q: Post-Training Quantization for Vision Transformers by Fisher Information Matrix Approximation

Less is More: Efficient Model Merging with Binary Task Switch

Lessons and Insights from a Unifying Study of Parameter-Efficient Fine-Tuning (PEFT) in Visual Recognition

Scaling Vision Pre-Training to 4K Resolution