papers

Publications (18)

cs.CV2026

MMSI-Bench: A Benchmark for Multi-Image Spatial Intelligence

Sihan Yang, Runsen Xu, Yiman Xie +10

Spatial intelligence is essential for multimodal large language models (MLLMs) operating in the complex physical world. Existing benchmarks, however, probe only single-image relati…

cs.CV2026

Multi-SpatialMLLM: Multi-Frame Spatial Understanding with Multi-Modal Large Language Models

Runsen Xu, Weiyao Wang, Hao Tang +5

Multi-modal large language models (MLLMs) have rapidly advanced in visual tasks, yet their spatial understanding remains limited to single images, leaving them ill-suited for physi…

cs.CV2023

EmbodiedScan: A Holistic Multi-Modal 3D Perception Suite Towards Embodied AI

Tai Wang, Xiaohan Mao, Chenming Zhu +11

In the realm of computer vision and robotics, embodied agents are expected to explore their environment and carry out human instructions. This necessitates the ability to fully und…

cs.CV2026

CO^3: Cooperative Unsupervised 3D Representation Learning for Autonomous Driving

Runjian Chen, Yao Mu, Runsen Xu +5

Unsupervised contrastive learning for indoor-scene point clouds has achieved great successes. However, unsupervised learning point clouds in outdoor scenes remains challenging beca…

cs.CV2025

GVLM: Geometry Grounded Vision Language Model with Unified 3D Reconstruction and Spatial Reasoning

Wenbo Hu, Jingli Lin, Yilin Long +7

Vision-Language Models (VLMs) still lack robustness in spatial intelligence, demonstrating poor performance on spatial understanding and reasoning tasks. We attribute this gap to t…

cs.CV2025

VFlowOpt: A Token Pruning Framework for LMMs with Visual Information Flow-Guided Optimization

Sihan Yang, Runsen Xu, Chenhang Cui +3

Large Multimodal Models (LMMs) excel in visual-language tasks by leveraging numerous visual tokens for fine-grained visual information, but this token redundancy results in signifi…

cs.CV2023

MV-JAR: Masked Voxel Jigsaw and Reconstruction for LiDAR-Based Self-Supervised Pre-Training

Runsen Xu, Tai Wang, Wenwei Zhang +4

This paper introduces the Masked Voxel Jigsaw and Reconstruction (MV-JAR) method for LiDAR-based self-supervised pre-training and a carefully designed data-efficient 3D object dete…

cs.CV2024

Grounded 3D-LLM with Referent Tokens

Yilun Chen, Shuai Yang, Haifeng Huang +5

Prior studies on 3D scene understanding have primarily developed specialized models for specific tasks or required task-specific fine-tuning. In this study, we propose Grounded 3D-…

cs.CV2023

Fine-Grained Cross-View Geo-Localization Using a Correlation-Aware Homography Estimator

Xiaolong Wang, Runsen Xu, Zuofan Cui +2

In this paper, we introduce a novel approach to fine-grained cross-view geo-localization. Our method aligns a warped ground image with a corresponding GPS-tagged satellite image co…

cs.CV2025

MMSI-Video-Bench: A Holistic Benchmark for Video-Based Spatial Intelligence

Jingli Lin, Runsen Xu, Shaohao Zhu +11

Spatial understanding over continuous visual input is crucial for MLLMs to evolve into general-purpose assistants in physical environments. Yet there is still no comprehensive benc…

cs.CV2024

Chat-Scene: Bridging 3D Scene and Large Language Models with Object Identifiers

Haifeng Huang, Yilun Chen, Zehan Wang +8

Recent advancements in 3D Large Language Models (LLMs) have demonstrated promising capabilities for 3D scene understanding. However, previous methods exhibit deficiencies in genera…

cs.CV2026

Learning Video Generation for Robotic Manipulation with Collaborative Trajectory Control

Xiao Fu, Xintao Wang, Xian Liu +5

Recent advances in video diffusion models shows promise for generating robotic decision-making data, with trajectory conditions further enabling fine-grained control. However, exis…

cs.CV2025

MMScan: A Multi-Modal 3D Scene Dataset with Hierarchical Grounded Language Annotations

Ruiyuan Lyu, Jingli Lin, Tai Wang +8

With the emergence of LLMs and their integration with other data modalities, multi-modal 3D perception attracts more attention due to its connectivity to the physical world and mak…

cs.CV2021

LIFE: Lighting Invariant Flow Estimation

Zhaoyang Huang, Xiaokun Pan, Runsen Xu +4

We tackle the problem of estimating flow between two images with large lighting variations. Recent learning-based flow estimation frameworks have shown remarkable performance on im…

cs.CV2024

VLM-Grounder: A VLM Agent for Zero-Shot 3D Visual Grounding

Runsen Xu, Zhiwei Huang, Tai Wang +3

3D visual grounding is crucial for robots, requiring integration of natural language and 3D scene understanding. Traditional methods depending on supervised learning with 3D point…

cs.CV2025

ChangingGrounding: 3D Visual Grounding in Changing Scenes

Miao Hu, Zhiwei Huang, Tai Wang +4

Real-world robots localize objects from natural-language instructions while scenes around them keep changing. Yet most of the existing 3D visual grounding (3DVG) method still assum…

cs.CV2024

PointLLM: Empowering Large Language Models to Understand Point Clouds

Runsen Xu, Xiaolong Wang, Tai Wang +3

The unprecedented advancements in Large Language Models (LLMs) have shown a profound impact on natural language processing but are yet to fully embrace the realm of 3D understandin…

cs.CV2025

OST-Bench: Evaluating the Capabilities of MLLMs in Online Spatio-temporal Scene Understanding

Jingli Lin, Chenming Zhu, Runsen Xu +4

Recent advances in multimodal large language models (MLLMs) have shown remarkable capabilities in integrating vision and language for complex reasoning. While most existing benchma…