Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Wang, Chunshi, Ye, Junliang, Yang, Yunhan, Li, Yang, Lin, Zizhuo, Zhu, Jun, Chen, Zhuo, Luo, Yawei, Guo, Chunchao
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2511.13647
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866914161482530816
author Wang, Chunshi
Ye, Junliang
Yang, Yunhan
Li, Yang
Lin, Zizhuo
Zhu, Jun
Chen, Zhuo
Luo, Yawei
Guo, Chunchao
author_facet Wang, Chunshi
Ye, Junliang
Yang, Yunhan
Li, Yang
Lin, Zizhuo
Zhu, Jun
Chen, Zhuo
Luo, Yawei
Guo, Chunchao
contents We introduce Part-X-MLLM, a native 3D multimodal large language model that unifies diverse 3D tasks by formulating them as programs in a structured, executable grammar. Given an RGB point cloud and a natural language prompt, our model autoregressively generates a single, coherent token sequence encoding part-level bounding boxes, semantic descriptions, and edit commands. This structured output serves as a versatile interface to drive downstream geometry-aware modules for part-based generation and editing. By decoupling the symbolic planning from the geometric synthesis, our approach allows any compatible geometry engine to be controlled through a single, language-native frontend. We pre-train a dual-encoder architecture to disentangle structure from semantics and instruction-tune the model on a large-scale, part-centric dataset. Experiments demonstrate that our model excels at producing high-quality, structured plans, enabling state-of-the-art performance in grounded Q\&A, compositional generation, and localized editing through one unified interface. Project page: https://chunshi.wang/Part-X-MLLM/
format Preprint
id arxiv_https___arxiv_org_abs_2511_13647
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Part-X-MLLM: Part-aware 3D Multimodal Large Language Model
Wang, Chunshi
Ye, Junliang
Yang, Yunhan
Li, Yang
Lin, Zizhuo
Zhu, Jun
Chen, Zhuo
Luo, Yawei
Guo, Chunchao
Computer Vision and Pattern Recognition
We introduce Part-X-MLLM, a native 3D multimodal large language model that unifies diverse 3D tasks by formulating them as programs in a structured, executable grammar. Given an RGB point cloud and a natural language prompt, our model autoregressively generates a single, coherent token sequence encoding part-level bounding boxes, semantic descriptions, and edit commands. This structured output serves as a versatile interface to drive downstream geometry-aware modules for part-based generation and editing. By decoupling the symbolic planning from the geometric synthesis, our approach allows any compatible geometry engine to be controlled through a single, language-native frontend. We pre-train a dual-encoder architecture to disentangle structure from semantics and instruction-tune the model on a large-scale, part-centric dataset. Experiments demonstrate that our model excels at producing high-quality, structured plans, enabling state-of-the-art performance in grounded Q\&A, compositional generation, and localized editing through one unified interface. Project page: https://chunshi.wang/Part-X-MLLM/
title Part-X-MLLM: Part-aware 3D Multimodal Large Language Model
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2511.13647