用替换式去噪预训练点云模型，提升结构学习能力与性能。
Point-RTD: Replaced Token Denoising for Pretraining Transformer Models on Point Clouds
- 通过替换点云令牌并重建，增强模型对结构信息的鲁棒性。
- 在ShapeNet上重建误差降低93%，测试集Chamfer Distance降低14倍以上。
- 收敛更快，分类准确率全面超越PointMAE，适合点云任务研究者。
预训练策略对基于Transformer的3D点云模型性能提升至关重要。本文提出Point-RTD（替换令牌去噪），一种新型预训练方法，通过污染-重建框架提升令牌鲁棒性。不同于传统掩码重建仅遮蔽数据片段，Point-RTD主动替换点云令牌，并采用判别器-生成器架构进行去噪，从而更有效地学习结构先验。在ShapeNet数据集上，Point-RTD相比PointMAE将重建误差降低超过93%，测试集Chamfer Distance降低超过14倍；同时在分类任务中收敛更快，在ShapeNet、ModelNet10和ModelNet40基准上均显著优于基线PointMAE。
原文摘要 · Abstract (English)
Pre-training strategies play a critical role in advancing the performance of transformer-based models for 3D point cloud tasks. In this paper, we introduce Point-RTD (Replaced Token Denoising), a novel pretraining strategy designed to improve token robustness through a corruption-reconstruction framework. Unlike traditional mask-based reconstruction tasks that hide data segments for later prediction, Point-RTD corrupts point cloud tokens and leverages a discriminator-generator architecture for denoising. This shift enables more effective learning of structural priors and significantly enhances model performance and efficiency. On the ShapeNet dataset, Point-RTD reduces reconstruction error by over 93% compared to PointMAE, and achieves more than 14x lower Chamfer Distance on the test set. Our method also converges faster and yields higher classification accuracy on ShapeNet, ModelNet10, and ModelNet40 benchmarks, clearly outperforming the baseline Point-MAE framework in every case.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。