Abstract
3D visual grounding aims to identify the target object within a 3D point cloud scene referred to by a natural language description. While previous works attempt to exploit the verbo-visual relation with proposed cross-modal transformers, unstructured natural utterances and scattered objects might lead to undesirable performances. In this paper, we introduce DOrA, a novel 3D visual grounding framework with Order-Aware referring. DOrA is designed to leverage Large Language Models (LLMs) to parse language description, suggesting a referential order of anchor objects. Such ordered anchor objects allow DOrA to update visual features and locate the target object during the grounding process. Experimental results on the NR3D and ScanRefer datasets demonstrate our superiority in both low-resource and full-data scenarios. In particular, DOrA surpasses current state-of-the-art frameworks by 9.3% and 7.8% grounding accuracy under 1% data and 10% data settings, respectively.
Abstract (translated)
3D视觉 grounded 旨在通过自然语言描述中的目标对象，在3D点云场景中确定目标对象。然而，以前的工作试图利用所提出的跨模态变换器利用动词-视觉关系，但无结构的自然语句和分散的对象可能会导致不良的性能。在本文中，我们引入了DOrA，一种新颖的3D视觉 grounded 框架，具有Order-Aware参考。DOrA旨在利用大型语言模型（LLMs）解析语言描述，建议锚对象之间的参照顺序。这样的有序锚对象允许DOrA在 grounding 过程中更新视觉特征并定位目标对象。在NR3D和ScanRefer数据集上的实验结果证实了我们在低资源和高资源场景中的卓越性。特别地，DOrA在1%数据和10%数据设置下的grounding准确度分别比现有最先进框架高9.3%和7.8%。
URL
https://arxiv.org/abs/2403.16539
