8 papers
DynFlowDrive: Flow-Based Dynamic World Modeling for Autonomous Driving
Xiaolu Liu, Yicong Li, Song Wang +3
Recently, world models have been incorporated into the autonomous driving systems to improve the planning reliability. Existing approaches typically predict future states through a…
VisionTrim: Unified Vision Token Compression for Training-Free MLLM Acceleration
Hanxun Yu, Wentong Li, Xuan Qu +3
Multimodal large language models (MLLMs) suffer from high computational costs due to excessive visual tokens, particularly in high-resolution and video-based scenarios. Existing to…
A Coarse-to-Fine Approach to Multi-Modality 3D Occupancy Grounding
Zhan Shi, Song Wang, Junbo Chen +1
Visual grounding aims to identify objects or regions in a scene based on natural language descriptions, essential for spatially aware perception in autonomous driving. However, exi…
Inst3D-LMM: Instance-Aware 3D Scene Understanding with Multi-modal Instruction Tuning
Hanxun Yu, Wentong Li, Song Wang +2
Despite encouraging progress in 3D scene understanding, it remains challenging to develop an effective Large Multi-modal Model (LMM) that is capable of understanding and reasoning…
PointLoRA: Low-Rank Adaptation with Token Selection for Point Cloud Learning
Song Wang, Xiaolu Liu, Lingdong Kong +6
Self-supervised representation learning for point cloud has demonstrated effectiveness in improving pre-trained model performance across diverse tasks. However, as pre-trained mode…
Offboard Occupancy Refinement with Hybrid Propagation for Autonomous Driving
Hao Shi, Song Wang, Jiaming Zhang +5
Vision-based occupancy prediction, also known as 3D Semantic Scene Completion (SSC), presents a significant challenge in computer vision. Previous methods, confined to onboard proc…