解决3D点云自监督学习中的位置信息泄露问题，提升语义表征质量。
Mitigating Positional Leakage in 3D Masked Autoencoders for Robust Representation Learning
- 设计可重校准的位置嵌入模块，抑制坐标信号过强影响。
- 引入门控位置接口，动态调节重建时位置信息的注入程度。
- 适合做3D点云表征学习的研究者和工业应用开发者。
掩码自编码已成为3D点云自监督学习的重要范式，在下游任务中表现优异。与2D版本不同，3D掩码自编码直接重建空间坐标，因此天然易受位置信息泄露影响。本文发现现有3D MAE框架的解码器过度依赖位置信息，削弱了语义表征学习，导致特征质量不佳。为此，我们提出MPL-MAE，一种缓解位置过依赖并增强编码器特征利用的掩码点学习框架。具体包括：一个重校准的位置嵌入模块，抑制度量主导的坐标信号同时保留几何拓扑；以及一个门控位置接口模块，动态调节重建过程中的位置注入。这些设计促进空间先验与语义特征的平衡交互，生成鲁棒且信息丰富的表示。大量下游任务实验表明，MPL-MAE持续取得有竞争力的性能，验证其有效性。代码已开源：https://github.com/yanx57/MPL-MAE。
原文摘要 · Abstract (English)
Masked autoencoding has emerged as a prominent paradigm for self-supervised learning on 3D point clouds, achieving competitive performance across downstream tasks. Unlike its 2D counterpart, 3D masked autoencoding directly reconstructs spatial coordinates, making it inherently susceptible to positional leakage. In this work, we identify that the decoder in existing 3D MAE frameworks tends to over-rely on positional information, which weakens semantic representation learning and leads to suboptimal feature quality. To address this issue, we propose MPL-MAE, a masked point learning framework that mitigates positional over-reliance while enhancing the utilization of encoder features. Specifically, we introduce a recalibrated positional embedding module that suppresses metric-dominant coordinate signals while preserving geometric topology, together with a gated positional interface module that dynamically regulates positional injection during reconstruction. These designs promote a more balanced interaction between spatial priors and semantic features, yielding robust and informative representations. Extensive experiments across downstream tasks demonstrate that MPL-MAE consistently achieves competitive performance, validating its effectiveness. Code is available at https://github.com/yanx57/MPL-MAE.
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。