Paper Overview
Research area: Computer Vision (CV) Authors: Shaohui Dai, Yansong Qu, You Shen Published: 2025-06-11 arXiv: 2506.08284
Summary
3D multimodal large language models (3D-MLLMs) have made progress in unified 3D scene understanding tasks, but existing models remain largely object-centric, limiting their ability to model fine-grained part structures.
This paper presents PAR3D, a unified part-aware 3D-MLLM framework that enables models to understand, reason about, and ground both objects and their parts in 3D scenes.
Key Contributions
- ScenePart dataset: A synthetic dataset introduced to support part-level 3D understanding and training.
- Part-Aware 3D Representation Learning: A representation learning approach designed to capture fine-grained part structures.
- Hierarchical segmentation query generation: Proposed for improved grounding of parts within scenes.
Results
Experiments demonstrate that the proposed method significantly improves performance on part-level question answering and referring segmentation tasks in 3D scenes.
Original Abstract
> Existing 3D-MLLMs remain largely object-centric, limiting their ability to model fine-grained part structures. We present PAR3D, a unified part-aware 3D-MLLM framework that enables models to understand, reason about, and ground both objects and their parts in 3D scenes. We introduce ScenePart dataset and develop Part-Aware 3D Representation Learning.
---
*Automatically collected on 2025-06-11*