CVPR 2025-Papers-about-Autonomous-Driving-and-Embodied-AI

Autonomous-Driving

(1) Occupancy

UniScene: Unified Occupancy-centric Driving Scene Generation

• paper: https://arxiv.org/pdf/2412.05435

• code: https://github.com/Arlo0o/UniScene-Unified-Occupancy-centric-Driving-Scene-Generation

GaussianFormer-2: Probabilistic Gaussian Superposition for Efficient 3D Occupancy Prediction

• paper: https://arxiv.org/abs/2412.04384

• code: https://github.com/huang-yh/GaussianFormer

(2) End-to-End Autonomous-Driving

Don't Shake the Wheel: Momentum-AwarePlanning in End-to-End Autonomous Driving

• paper: https://arxiv.org/pdf/2503.03125

• code: https://github.com/adept-thu/MomAD

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

• paper: https://arxiv.org/pdf/2411.15139

• code: https://github.com/hustvl/DiffusionDrive

(3) Multi-modal Fusion

V2X - R: Cooperative LiDAR - 4D Radar Fusion with Denoising Diffusion for 3D Object Detection

• paper: https://arxiv.org/pdf/2411.08402

• code: https://github.com/ylwhxht/V2X-R

(4) Vision Language Model

OmniDrive: A Holistic Vision - Language Dataset for Autonomous Driving with counter Factual Reasoning

• paper: https://arxiv.org/pdf/2405.01533

• code: https://github.com/NVlabs/OmniDrive

(5) World Model

DriveDreamer4D: World Models Are Effective Data Machines for 4D Driving Scene

• paper: https://arxiv.org/pdf/2410.13571

• code: https://github.com/GigaAI-research/DriveDreamer4D

ReconDreamer: Crafting World Models for Driving Scene Reconstruction via Online Restoration

• paper: https://arxiv.org/pdf/2411.19548

• code: https://github.com/GigaAI-research/ReconDreamer

(6) Closed Loop

DrivingSphere: Building a High-fidelity 4D World for Closed-loop Simulation

• paper: https://arxiv.org/abs/2411.11252

• code: https://github.com/yanty123/DrivingSphere

(7) Lane Detection

Rethinking Lanes and Points in complex Scenarios for Monocular 3D Lane Detection

• paper: https://arxiv.org/abs/2503.06237

(8) Motion Prediction

ModeSeq: Taming Sparse Multimodal Motion Prediction with Sequential Mode Modeling

• paper: https://arxiv.org/pdf/2411.11911

Embodied-AI

(1) Diffusion Model

DexHandDiff: Interaction - aware Diffusion Planning for Adaptive Dexterous Manipulation

• paper: https://arxiv.org/abs/2411.18562

(2) 3D Semantic Flow

G3Flow: Generative 3D Semantic Flow for Pose-aware and Generalizable Object Manipulation

• paper: https://arxiv.org/abs/2411.18369
• code: https://github.com/TianxingChen/G3Flow

(3) Dual-Arm Robot

RoboTwin: Dual-Arm Robot Benchmark with Generative Digital Twins

• paper: https://arxiv.org/abs/2409.02920
• code: https://github.com/TianxingChen/RoboTwin?tab=readme-ov-file#robotwin-dual-arm-robot-benchmark-with-generative-digital-twins

Towards Long-Horizon Vision-Language Navigation: Platform, Benchmark and Method

• paper: https://arxiv.org/pdf/2412.09082
• code: https://github.com/HCPLab-SYSU/LH-VLN

(4) Visual Language Model

Code-as-Monitor: Constraint-aware Visual Programming for Reactive and Proactive Robotic Failure Detection

• paper: https://arxiv.org/abs/2412.04455

PhysVLM: Enabling Visual Language Models to Understand Robotic Physical Reachability

--paper: https://arxiv.org/abs/2503.08481

Others

Mamba4D: Efficient 4D Point Cloud Video Understanding with Disentangled Spatial-Temporal State Space Models

• paper: https://arxiv.org/abs/2405.14338

• code: https://github.com/IRMVLab/Mamba4D

Timestep Embedding Tells: It's Time to Cache for Video Diffusion Model

• paper: https://arxiv.org/abs/2411.19108

• code: https://github.com/ali-vilab/TeaCache

DynRefer: Delving into Region-level Multi-modality Tasks via Dynamic Resolution

• paper: https://arxiv.org/abs/2405.16071

• code: https://github.com/callsys/DynRefer

OmniDocBench: Benchmarking Diverse PDF Document Parsing with Comprehensive Annotations

• paper: https://arxiv.org/pdf/2412.07626

• code: https://arxiv.org/pdf/2412.07626

VTON360: High-Fidelity Virtual Try-On from Any Viewing Direction

• paper: https://arxiv.org/abs/2503.12165

• code: https://github.com/scnuhealthy/VTON360

Cross-modal Causal Relation Alignment for Video Question Grounding

-- paper: https://arxiv.org/abs/2503.07635

-- code: https://github.com/WissingChen/CRA-GQA

DSPNet: Dual-vision Scene Perception for Robust 3D Question Answering

-- paper: https://arxiv.org/abs/2503.03190

-- code: https://github.com/LZ-CH/DSPNet

Reproducible Vision-Language Models Meet Concepts Out of Pre-Training

LLM-driven Multimodal and Multi-Identity Listening Head Generation

DAGSM: Disentangled Avatar Generation with GS-enhanced Mesh

-- paper: https://arxiv.org/abs/2411.15205

HiRes-LLaVA: Restoring Fragmentation Input in High-Resolution Large Vision-Language Models

-- paper: https://arxiv.org/abs/2407.08706

FireEdit: Fine-grained Instruction-based Image Editing via Region-aware Vision Language Model

-- paper: https://arxiv.org/abs/2503.19839

-- code: https://zjgans.github.io/fireedit.github.io/

EMOVA: Empowering Language Models to See, Hear and Speak with Vivid Emotions

-- paper: https://arxiv.org/abs/2409.18042

-- code: https://emova-anonymous.github.io/

PS-Diffusion: Photorealistic Subject-Driven Image Editing with Disentangled Control and Attention

Boosting the Dual-Stream Architecture in Ultra-High Resolution Segmentation with Resolution-Biased Uncertainty Estimation

No Pains, More Gains: Recycling Sub-Salient Patches for Efficient High-Resolution Image Recognition

Empowering Large Language Models with 3D Situation Awareness

-- paper: https://arxiv.org/abs/2503.23024

Rethinking Query-based Transformer for Continual Image Segmentation

Eval3D: Interpretable and Fine-grained Evaluation for 3D Generation

• paper: http://arxiv.org/pdf/2504.18509v1

• code: https://eval3d.git

Seeing Soundscapes: Audio-Visual Generation and Separation from Soundscapes Using Audio-Visual Separator

• paper: http://arxiv.org/pdf/2504.18283v1

SSL4Eco: A Global Seasonal Dataset for Geospatial Foundation Models in Ecology

• paper: http://arxiv.org/pdf/2504.18256v1

• code: https://github.com/PlekhanovaElena/ssl4eco

BiasBench: A reproducible benchmark for tuning the biases of event cameras

• paper: http://arxiv.org/pdf/2504.18235v1

Enhancing Privacy-Utility Trade-offs to Mitigate Memorization in Diffusion Models

• paper: http://arxiv.org/pdf/2504.18032v1

• code: https://chenchen-usyd.git

Dynamic Camera Poses and Where to Find Them

• paper: http://arxiv.org/pdf/2504.17788v1

• code: https://research.nvidia.com/labs/dir/dynpose-100k

PICO: Reconstructing 3D People In Contact with Objects

• paper: http://arxiv.org/pdf/2504.17695v1

• code: https://pico.is.tue.mpg.de

VEU-Bench: Towards Comprehensive Understanding of Video Editing

• paper: http://arxiv.org/pdf/2504.17828v1

Dual Prompting Image Restoration with Diffusion Transformers

• paper: http://arxiv.org/pdf/2504.17825v1

SemanticSugarBeets: A Multi-Task Framework and Dataset for Inspecting Harvest and Storage Characteristics of Sugar Beets

• paper: http://arxiv.org/pdf/2504.16684v1

• code: https://github.com/semanticsugarbeets/semanticsugarbeets

PRaDA: Projective Radial Distortion Averaging

• paper: http://arxiv.org/pdf/2504.16499v1

CLOC: Contrastive Learning for Ordinal Classification with Multi-Margin N-pair Loss

• paper: http://arxiv.org/pdf/2504.17813v1

LiveCC: Learning Video LLM with Streaming Speech Transcription at Scale

• paper: http://arxiv.org/pdf/2504.16030v1

• code: https://showlab.git

PointLoRA: Low-Rank Adaptation with Token Selection for Point Cloud Learning

• paper: http://arxiv.org/pdf/2504.16023v1

• code: https://github.com/songw-zju/PointLoRA

RepNet-VSR: Reparameterizable Architecture for High-Fidelity Video Super-Resolution

• paper: http://arxiv.org/pdf/2504.15649v1

MirrorVerse: Pushing Diffusion Models to Realistically Reflect the World

• paper: http://arxiv.org/pdf/2504.15397v1

• code: https://mirror-verse.git

Plug-and-Play Versatile Compressed Video Enhancement

• paper: http://arxiv.org/pdf/2504.15380v1

• code: https://huimin-zeng.git

Acquire and then Adapt: Squeezing out Text-to-Image Model for Image Restoration

• paper: http://arxiv.org/pdf/2504.15159v1

Improving Sound Source Localization with Joint Slot Attention on Image and Audio

• paper: http://arxiv.org/pdf/2504.15118v1

NTIRE 2025 Challenge on Short-form UGC Video Quality Assessment and Enhancement: KwaiSR Dataset and Study

• paper: http://arxiv.org/pdf/2504.15003v1

• code: https://lixinustc.git

DyFo: A Training-Free Dynamic Focus Visual Search for Enhancing LMMs in Fine-Grained Visual Understanding

• paper: http://arxiv.org/pdf/2504.14920v1

• code: https://github.com/PKU-ICST-MIPL/DyFo_CVPR2025

Other Unfiled papers

LiveCC: Learning Video LLM with Streaming Speech Transcription at Scale

• paper: https://arxiv.org/pdf/2504.16030

An LLM-enabled Multi-Agent Autonomous Mechatronics Design Framework

• paper: https://arxiv.org/pdf/2504.14681

CheXWorld: Exploring Image World Modeling for Radiograph Representation Learning

• paper: https://arxiv.org/pdf/2504.13820

EchoWorld: Learning Motion-Aware World Models for Echocardiography Probe Guidance

• paper: https://arxiv.org/pdf/2504.13065

RoboTwin: Dual-Arm Robot Benchmark with Generative Digital Twins

• paper: https://arxiv.org/pdf/2504.13059

Rethinking Temporal Fusion with a Unified Gradient Descent View for 3D Semantic Occupancy Prediction

• paper: https://arxiv.org/pdf/2504.12959

ZeroGrasp: Zero-Shot Shape Reconstruction Enabled Robotic Grasping

• paper: https://arxiv.org/pdf/2504.10857

CleanMAP: Distilling Multimodal LLMs for Confidence-Driven Crowdsourced HD Map Updates

• paper: https://arxiv.org/pdf/2504.10738

Foundation Models for Remote Sensing: An Analysis of MLLMs for Object Localization

• paper: https://arxiv.org/pdf/2504.10727

SilVar-Med: A Speech-Driven Visual Language Model for Explainable Abnormality Detection in Medical Imaging

• paper: https://arxiv.org/pdf/2504.10642

Ges3ViG: Incorporating Pointing Gestures into Language-Based 3D Visual Grounding for Embodied Reference Understanding

• paper: https://arxiv.org/pdf/2504.09623

Gen3DEval: Using vLLMs for Automatic Evaluation of Generated 3D Objects

• paper: https://arxiv.org/pdf/2504.08125

MESA: Text-Driven Terrain Generation Using Latent Diffusion and Global Copernicus Data

• paper: https://arxiv.org/pdf/2504.07210

Neural Motion Simulator: Pushing the Limit of World Models in Reinforcement Learning

• paper: https://arxiv.org/pdf/2504.07095

Two by Two: Learning Multi-Task Pairwise Objects Assembly for Generalizable Robot Manipulation

• paper: https://arxiv.org/pdf/2504.06961

From Broadcast to Minimap: Achieving State-of-the-Art SoccerNet Game State Reconstruction

• paper: https://arxiv.org/pdf/2504.06357

OCC-MLLM-CoT-Alpha: Towards Multi-stage Occlusion Recognition Based on Large Language Models via 3D-Aware Supervision and Chain-of-Thoughts Guidance

• paper: https://arxiv.org/pdf/2504.04781

Data Scaling Laws for End-to-End Autonomous Driving

• paper: https://arxiv.org/pdf/2504.04338

Decision SpikeFormer: Spike-Driven Transformer for Decision Making

• paper: https://arxiv.org/pdf/2504.03800

Shape My Moves: Text-Driven Shape-Aware Synthesis of Human Motions

• paper: https://arxiv.org/pdf/2504.03639

ZFusion: An Effective Fuser of Camera and 4D Radar for 3D Object Perception in Autonomous Driving

• paper: https://arxiv.org/pdf/2504.03438

AI Hiring with LLMs: A Context-Aware and Explainable Multi-Agent Framework for Resume Screening

• paper: https://arxiv.org/pdf/2504.02870

Exploration-Driven Generative Interactive Environments

• paper: https://arxiv.org/pdf/2504.02515

Think Small, Act Big: Primitive Prompt Learning for Lifelong Robot Manipulation

• paper: https://arxiv.org/pdf/2504.00420

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

• paper: https://arxiv.org/pdf/2504.00379

Chapter-Llama: Efficient Chaptering in Hour-Long Videos with LLMs

• paper: https://arxiv.org/pdf/2504.00072

Scenario Dreamer: Vectorized Latent Diffusion for Generating Driving Simulation Environments

• paper: https://arxiv.org/pdf/2503.22496

Protecting Your Video Content: Disrupting Automated Video-based LLM Annotations

• paper: https://arxiv.org/pdf/2503.21824

CoMapGS: Covisibility Map-based Gaussian Splatting for Sparse Novel View Synthesis

• paper: https://arxiv.org/pdf/2503.20998

BioX-CPath: Biologically-driven Explainable Diagnostics for Multistain IHC Computational Pathology

• paper: https://arxiv.org/pdf/2503.20880

CoLLM: A Large Language Model for Composed Image Retrieval

• paper: https://arxiv.org/pdf/2503.19910

Attention IoU: Examining Biases in CelebA using Attention Maps

• paper: https://arxiv.org/pdf/2503.19846

AudCast: Audio-Driven Human Video Generation by Cascaded Diffusion Transformers

• paper: https://arxiv.org/pdf/2503.19824

Teller: Real-Time Streaming Audio-Driven Portrait Animation with Autoregressive Motion Generation

• paper: https://arxiv.org/pdf/2503.18429

HiLoTs: High-Low Temporal Sensitive Representation Learning for Semi-Supervised LiDAR Segmentation in Autonomous Driving

• paper: https://arxiv.org/pdf/2503.17752

CountLLM: Towards Generalizable Repetitive Action Counting via Large Language Model

• paper: https://arxiv.org/pdf/2503.17690

Missing Target-Relevant Information Prediction with World Model for Accurate Zero-Shot Composed Image Retrieval

• paper: https://arxiv.org/pdf/2503.17109

Vision-Language Gradient Descent-driven All-in-One Deep Unfolding Networks

• paper: https://arxiv.org/pdf/2503.16930

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

• paper: https://arxiv.org/pdf/2503.15871

Generating Multimodal Driving Scenes via Next-Scene Prediction

• paper: https://arxiv.org/pdf/2503.14945

When Domain Generalization meets Generalized Category Discovery: An Adaptive Task-Arithmetic Driven Approach

• paper: https://arxiv.org/pdf/2503.14897

Bridging Past and Future: End-to-End Autonomous Driving with Historical Prediction and Planning

• paper: https://arxiv.org/pdf/2503.14182

MP-GUI: Modality Perception with MLLMs for GUI Understanding

• paper: https://arxiv.org/pdf/2503.14021

DIFFVSGG: Diffusion-Driven Online Video Scene Graph Generation

• paper: https://arxiv.org/pdf/2503.13957

Conformal Prediction and MLLM aided Uncertainty Quantification in Scene Graph Generation

• paper: https://arxiv.org/pdf/2503.13947

SALAD: Skeleton-aware Latent Diffusion for Text-driven Motion Generation and Editing

• paper: https://arxiv.org/pdf/2503.13836

Omnia de EgoTempo: Benchmarking Temporal Understanding of Multi-Modal LLMs in Egocentric Videos

• paper: https://arxiv.org/pdf/2503.13646

Efficient Motion-Aware Video MLLM

• paper: https://arxiv.org/pdf/2503.13016

V-Stylist: Video Stylization via Collaboration and Reflection of MLLM Agents

• paper: https://arxiv.org/pdf/2503.12077

Minding Fuzzy Regions: A Data-driven Alternating Learning Paradigm for Stable Lesion Segmentation

• paper: https://arxiv.org/pdf/2503.11140

DriveGEN: Generalized and Robust 3D Detection in Driving via Controllable Text-to-Image Diffusion Generation

• paper: https://arxiv.org/pdf/2503.11122

Spatial-Temporal Graph Diffusion Policy with Kinematic Modeling for Bimanual Robotic Manipulation

• paper: https://arxiv.org/pdf/2503.10743

GaussHDR: High Dynamic Range Gaussian Splatting via Learning Unified 3D and 2D Local Tone Mapping

• paper: https://arxiv.org/pdf/2503.10143

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

• paper: https://arxiv.org/pdf/2503.09594

Out-of-Distribution Segmentation in Autonomous Driving: Problems and State of the Art

• paper: https://arxiv.org/pdf/2503.08695

DexGrasp Anything: Towards Universal Robotic Dexterous Grasping with Physics Awareness

• paper: https://arxiv.org/pdf/2503.08257

VidBot: Learning Generalizable 3D Actions from In-the-Wild 2D Human Videos for Zero-Shot Robotic Manipulation

• paper: https://arxiv.org/pdf/2503.07135

A Data-Centric Revisit of Pre-Trained Vision Models for Robot Learning

• paper: https://arxiv.org/pdf/2503.06960

Denoising Functional Maps: Diffusion Models for Shape Correspondence

• paper: https://arxiv.org/pdf/2503.01845

KeyFace: Expressive Audio-Driven Facial Animation for Long Sequences via KeyFrame Interpolation

• paper: https://arxiv.org/pdf/2503.01715

Order-Robust Class Incremental Learning: Graph-Driven Dynamic Similarity Grouping

• paper: https://arxiv.org/pdf/2502.20032

CarPlanner: Consistent Auto-regressive Trajectory Planning for Large-scale Reinforcement Learning in Autonomous Driving

• paper: https://arxiv.org/pdf/2502.19908

Transfer Your Perspective: Controllable 3D Generation from Any Viewpoint in a Driving Scene

• paper: https://arxiv.org/pdf/2502.06682

Universal Actions for Enhanced Embodied Foundation Models

• paper: https://arxiv.org/pdf/2501.10105

OVO-Bench: How Far is Your Video-LLMs from Real-World Online Video Understanding?

• paper: https://arxiv.org/pdf/2501.05510

MotionMap: Representing Multimodality in Human Pose Forecasting

• paper: https://arxiv.org/pdf/2412.18883

Adapter Merging with Centroid Prototype Mapping for Scalable Class-Incremental Learning

• paper: https://arxiv.org/pdf/2412.18219

GME: Improving Universal Multimodal Retrieval by Multimodal LLMs

• paper: https://arxiv.org/pdf/2412.16855

Empowering LLMs to Understand and Generate Complex Vector Graphics

• paper: https://arxiv.org/pdf/2412.11102

StyleStudio: Text-Driven Style Transfer with Selective Control of Style Elements

• paper: https://arxiv.org/pdf/2412.08503

DiffSensei: Bridging Multi-Modal LLMs and Diffusion Models for Customized Manga Generation

• paper: https://arxiv.org/pdf/2412.07589

UniScene: Unified Occupancy-centric Driving Scene Generation

• paper: https://arxiv.org/pdf/2412.05435

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

• paper: https://arxiv.org/pdf/2412.04464

Code-as-Monitor: Constraint-aware Visual Programming for Reactive and Proactive Robotic Failure Detection

• paper: https://arxiv.org/pdf/2412.04455

Navigation World Models

• paper: https://arxiv.org/pdf/2412.03572

Task-driven Image Fusion with Learnable Fusion Loss

• paper: https://arxiv.org/pdf/2412.03240

UniGraspTransformer: Simplified Policy Distillation for Scalable Dexterous Robotic Grasping

• paper: https://arxiv.org/pdf/2412.02699

Video-3D LLM: Learning Position-Aware Video Representation for 3D Scene Understanding

• paper: https://arxiv.org/pdf/2412.00493

Immune: Improving Safety Against Jailbreaks in Multi-modal LLMs via Inference-Time Alignment

• paper: https://arxiv.org/pdf/2411.18688

CityWalker: Learning Embodied Urban Navigation from Web-Scale Videos

• paper: https://arxiv.org/pdf/2411.17820

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

• paper: https://arxiv.org/pdf/2411.17646

Augmenting Multimodal LLMs with Self-Reflective Tokens for Knowledge-based Visual Question Answering

• paper: https://arxiv.org/pdf/2411.16863

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

• paper: https://arxiv.org/pdf/2411.16537

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

• paper: https://arxiv.org/pdf/2411.15139

Voxel-Aggregated Feature Synthesis: Efficient Dense Mapping for Simulated 3D Reasoning

• paper: https://arxiv.org/pdf/2411.10616

Decoupling Fine Detail and Global Geometry for Compressed Depth Map Super-Resolution

• paper: https://arxiv.org/pdf/2411.03239

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

• paper: https://arxiv.org/pdf/2410.23780

Style-Editor: Text-driven object-centric style editing

• paper: https://arxiv.org/pdf/2408.08461

3D-MVP: 3D Multiview Pretraining for Robotic Manipulation

• paper: https://arxiv.org/pdf/2406.18158

Mitigating the Human-Robot Domain Discrepancy in Visual Pre-training for Robotic Manipulation

• paper: https://arxiv.org/pdf/2406.14235

3D-GRAND: A Million-Scale Dataset for 3D-LLMs with Better Grounding and Less Hallucination

• paper: https://arxiv.org/pdf/2406.05132

VideoTree: Adaptive Tree-based Video Representation for LLM Reasoning on Long Videos

• paper: https://arxiv.org/pdf/2405.19209