arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.20409cs.CVcs.LG
收藏

CLIPoint3D: 基于语言的少样本无监督3D点云领域适应

CLIPoint3D: Language-Grounded Few-Shot Unsupervised 3D Point Cloud Domain Adaptation

University of Trento(特伦托大学);MDSR Labs Adobe(Adobe MDSR实验室);IIT Bombay(印度理工学院班加罗尔分校);Fondazione Bruno Kessler(布鲁诺·科斯勒基金会)

Mainak Singha, Sarthak Mehrotra, Paolo Casari, Subhasis Chaudhuri, Elisa Ricci, Biplab Banerjee

AI总结:

本文提出CLIPoint3D,首个基于CLIP的3D点云无监督领域适应框架,通过知识驱动的提示调优和熵引导视图采样策略,实现跨领域3D点云适应,实验表明在PointDA-10和GraspNetPC-10基准上性能提升3-16%。

摘要:

近期的视觉-语言模型(如CLIP)展示了出色的跨模态推理能力,扩展至3D感知。然而,这些模型在领域迁移时表现脆弱,尤其在从合成到真实点云适应时。传统3D领域适应方法依赖于重参数化编码器,虽然准确率高但效率低。我们引入CLIPoint3D,首个基于CLIP的3D点云少样本无监督领域适应框架。我们的方法将3D样本投影到多个深度图,并利用冻结的CLIP主干网络,通过知识驱动的提示调优方案进行优化,该方案整合了高层语言先验与轻量3D编码器的几何线索。为有效适应任务特定特征,我们对CLIP编码器应用参数高效微调,并设计熵引导的视图采样策略以选择自信投影。此外,最优传输基于对齐损失和不确定性感知原型对齐损失协同解决源-目标分布差距,同时保持类别分离性。在PointDA-10和GraspNetPC-10基准上的大量实验表明,CLIPoint3D在CLIP基线和传统编码器基线上均实现了3-16%的准确率提升。项目页面:https://sarthakm320.github.io/CLIPoint3D。

英文摘要:

Recent vision-language models (VLMs) such as CLIP demonstrate impressive cross-modal reasoning, extending beyond images to 3D perception. Yet, these models remain fragile under domain shifts, especially when adapting from synthetic to real-world point clouds. Conventional 3D domain adaptation approaches rely on heavy trainable encoders, yielding strong accuracy but at the cost of efficiency. We introduce CLIPoint3D, the first framework for few-shot unsupervised 3D point cloud domain adaptation built upon CLIP. Our approach projects 3D samples into multiple depth maps and exploits the frozen CLIP backbone, refined through a knowledge-driven prompt tuning scheme that integrates high-level language priors with geometric cues from a lightweight 3D encoder. To adapt task-specific features effectively, we apply parameter-efficient fine-tuning to CLIP's encoders and design an entropy-guided view sampling strategy for selecting confident projections. Furthermore, an optimal transport-based alignment loss and an uncertainty-aware prototype alignment loss collaboratively bridge source-target distribution gaps while maintaining class separability. Extensive experiments on PointDA-10 and GraspNetPC-10 benchmarks show that CLIPoint3D achieves consistent 3-16% accuracy gains over both CLIP-based and conventional encoder-based baselines. Project page: https://sarthakm320.github.io/CLIPoint3D.

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。

↑