EDA: 面向3D视觉定位的显式文本解耦与稠密对齐
EDA: Explicit Text-Decoupling and Dense Alignment for 3D Visual Grounding
AI总结：
针对现有3D视觉定位方法丢失词级信息或忽略属性的问题，提出EDA框架，通过文本解耦和双损失监督的稠密对齐实现性能提升，还提出新任务验证模型能力并取得领先结果。
摘要：
3D视觉定位旨在点云中找到由包含丰富语义线索的自由形式自然语言描述所提及的物体。然而，现有方法要么提取耦合所有单词的句子级特征，要么更关注物体名称，这会丢失单词级信息或忽略其他属性。为缓解这些问题，我们提出EDA，即显式解耦句子中的文本属性，并在这种细粒度语言与点云物体之间进行稠密对齐。具体而言，我们首先提出一个文本解耦模块，为每个语义成分生成文本特征。随后，我们设计两种损失来监督两种模态间的稠密匹配：位置对齐损失和语义对齐损失。在此基础上，我们进一步引入一项新的视觉定位任务——定位不含物体名称的物体，该任务可全面评估模型的稠密对齐能力。通过实验，我们在两个广泛使用的3D视觉定位数据集ScanRefer和SR3D/NR3D上取得了最优性能，并且在我们新提出的任务上获得了绝对领先的表现。源代码可在https://github.com/yanmin-wu/EDA获取。
英文摘要：
3D visual grounding aims to find the object within point clouds mentioned by free-form natural language descriptions with rich semantic cues. However, existing methods either extract the sentence-level features coupling all words or focus more on object names, which would lose the word-level information or neglect other attributes. To alleviate these issues, we present EDA that Explicitly Decouples the textual attributes in a sentence and conducts Dense Alignment between such fine-grained language and point cloud objects. Specifically, we first propose a text decoupling module to produce textual features for every semantic component. Then, we design two losses to supervise the dense matching between two modalities: position alignment loss and semantic alignment loss. On top of that, we further introduce a new visual grounding task, locating objects without object names, which can thoroughly evaluate the model's dense alignment capacity. Through experiments, we achieve state-of-the-art performance on two widely-adopted 3D visual grounding datasets, ScanRefer and SR3D/NR3D, and obtain absolute leadership on our newly-proposed task. The source code is available at https://github.com/yanmin-wu/EDA.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
