Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2602.03890 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918529344733184 |
|---|---|
| author | Zhang, Xindan Yan, Weilong Shi, Yufei Qiu, Xuerui He, Tao Li, Ying Li, Ming Fan, Hehe |
| author_facet | Zhang, Xindan Yan, Weilong Shi, Yufei Qiu, Xuerui He, Tao Li, Ying Li, Ming Fan, Hehe |
| contents | Point clouds provide a compact and expressive representation of 3D objects, and have recently been integrated into multimodal large language models (MLLMs). However, existing methods primarily focus on static objects, while understanding dynamic point cloud sequences remains largely unexplored. This limitation is mainly caused by the lack of large-scale cross-modal datasets and the difficulty of modeling motions in spatio-temporal contexts. To bridge this gap, we present 4DPC$^2$hat, the first MLLM tailored for dynamic point cloud understanding. To this end, we construct a large-scale cross-modal dataset 4DPC$^2$hat-200K via a meticulous two-stage pipeline consisting of topology-consistent 4D point construction and two-level captioning. The dataset contains over 44K dynamic object sequences, 700K point cloud frames, and 200K curated question-answer (QA) pairs, supporting inquiries about counting, temporal relationship, action, spatial relationship, and appearance. At the core of the framework, we introduce a Mamba-enhanced temporal reasoning MLLM to capture long-range dependencies and dynamic patterns among a point cloud sequence. Furthermore, we propose a failure-aware bootstrapping learning strategy that iteratively identifies model deficiencies and generates targeted QA supervision to continuously strengthen corresponding reasoning capabilities. Extensive experiments demonstrate that our 4DPC$^2$hat significantly improves action understanding and temporal reasoning compared with existing models, establishing a strong foundation for 4D dynamic point cloud understanding. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_03890 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | 4DPC$^2$hat: Towards Dynamic Point Cloud Understanding with Failure-Aware Bootstrapping Zhang, Xindan Yan, Weilong Shi, Yufei Qiu, Xuerui He, Tao Li, Ying Li, Ming Fan, Hehe Computer Vision and Pattern Recognition Point clouds provide a compact and expressive representation of 3D objects, and have recently been integrated into multimodal large language models (MLLMs). However, existing methods primarily focus on static objects, while understanding dynamic point cloud sequences remains largely unexplored. This limitation is mainly caused by the lack of large-scale cross-modal datasets and the difficulty of modeling motions in spatio-temporal contexts. To bridge this gap, we present 4DPC$^2$hat, the first MLLM tailored for dynamic point cloud understanding. To this end, we construct a large-scale cross-modal dataset 4DPC$^2$hat-200K via a meticulous two-stage pipeline consisting of topology-consistent 4D point construction and two-level captioning. The dataset contains over 44K dynamic object sequences, 700K point cloud frames, and 200K curated question-answer (QA) pairs, supporting inquiries about counting, temporal relationship, action, spatial relationship, and appearance. At the core of the framework, we introduce a Mamba-enhanced temporal reasoning MLLM to capture long-range dependencies and dynamic patterns among a point cloud sequence. Furthermore, we propose a failure-aware bootstrapping learning strategy that iteratively identifies model deficiencies and generates targeted QA supervision to continuously strengthen corresponding reasoning capabilities. Extensive experiments demonstrate that our 4DPC$^2$hat significantly improves action understanding and temporal reasoning compared with existing models, establishing a strong foundation for 4D dynamic point cloud understanding. |
| title | 4DPC$^2$hat: Towards Dynamic Point Cloud Understanding with Failure-Aware Bootstrapping |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2602.03890 |