用于鲁棒点云分类的点选择微调框架
Point-Selection Fine-Tuning Framework for Robust Point Cloud Classification
- Guangdong Laboratory of Artificial Intelligence and Digital Economy (Shenzhen)（广东省人工智能与数字经济实验室（深圳））
- Shenzhen University（深圳大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
研究针对噪声和损坏点降低点云识别性能的问题，提出PSFT点选择微调框架，通过估计点影响、生成提示令牌、附加特征过滤器等操作，在保持参数高效的同时提高鲁棒性，实验证明其有效性。
AI中文摘要：
噪声和损坏的点会严重降低点云识别性能，尤其是在具有挑战性的损坏设置下。3D预训练模型的完全微调可能会放大异常值的影响并覆盖预训练期间学到的鲁棒性先验，而简单的参数高效适应对损坏的令牌仍然敏感。为解决此问题，我们提出了PSFT，一种点选择微调框架，在保持参数高效的同时提高鲁棒性。PSFT首先从预池化特征估计逐点影响，并自适应保留影响最小的点以抑制异常值。基于所选子集，提示生成分支预测逐层提示令牌并将其注入冻结的主干以进行轻量级下游适应。为进一步减轻选择后的残余噪声，我们附加了一个带有瓶颈MLP变换和Beta门控残差混合的轻量级特征过滤器，在预测前细化补丁令牌表示。大量实验表明，PSFT在所有测试的3D预训练主干上持续降低ModelNet-C和ModelNet40-C上的损坏错误，同时在评估的调整策略中使用ULIP-2和Uni3D-B获得最强的ScanObjectNN-C结果。
英文摘要：
Noisy and corrupted points can substantially degrade point cloud recognition performance, especially under challenging corruption settings. In particular, full fine-tuning of 3D pre-trained models may amplify the influence of outliers and overwrite robustness priors learned during pre-training, while naive parameter-efficient adaptation remains sensitive to corrupted tokens. To address this issue, we propose PSFT, a point-selection fine-tuning framework that improves robustness while remaining parameter-efficient. PSFT first estimates point-wise influence from pre-pooling features and adaptively retains minimally influential points to suppress outliers. Based on the selected subset, a prompt generation branch predicts layer-wise prompt tokens and injects them into a frozen backbone for lightweight downstream adaptation. To further mitigate residual noise after selection, we append a lightweight feature filter with bottleneck MLP transformation and Beta-gated residual blending to refine patch-token representations before prediction. Extensive experiments show that PSFT consistently reduces corruption error on ModelNet-C and ModelNet40-C across all tested 3D pre-trained backbones, while achieving the strongest ScanObjectNN-C results with ULIP-2 and Uni3D-B among the evaluated tuning strategies. Our implementation can be found at https://github.com/CVChMA/PSFT/tree/master.
