长尾3D点云数据集蒸馏
Long-Tailed 3D Point Cloud Dataset Distillation
浏览论文内容
中文总结 AI 辅助
本研究针对长尾3D点云数据集蒸馏问题，提出含自适应合成预算分配与3D长尾分布匹配模块的框架，在ShapeNet55上使分类准确率较SOTA提升7.0个百分点。
中文摘要 AI 辅助
数据集蒸馏可将大规模数据集压缩为紧凑的合成集，同时保留其训练效用，从而实现高效的3D点云训练。当前的点云数据集蒸馏方法仅处理几何与表示层面的挑战，却忽略了点云数据集中普遍存在的分布不平衡问题——其训练集与测试集均遵循长尾类分布。据我们所知，本研究是首个针对长尾点云数据集蒸馏的研究。我们的框架未仅聚焦于几何与表示属性，也未仅构建类平衡的合成集，而是通过两个核心模块明确考虑长尾类分布：首先，我们设计自适应合成预算分配（Adaptive Synthetic Budgeting），依据类样本数量与额外合成样本的预期收益分配各类别的合成预算；在分配的预算基础上，我们进一步设计3D长尾分布匹配（3D Long-Tailed Distribution Matching），通过全局-局部特征对齐与先验感知监督优化合成点云。前者保留全局类分布与多样的类内结构，后者提供依赖类别的专家监督，确保尾类样本可识别，同时维持多样的头类模式。大量实验验证了我们方法的有效性，在ShapeNet55数据集上，相比现有最优方法，分类准确率提升了7.0个百分点。
英文摘要
Dataset distillation compresses large-scale datasets into compact synthetic sets while preserving their training utility, enabling efficient 3D point cloud training. Current point cloud dataset distillation methods only tackle geometric and representation challenges while ignoring the distributional imbalance prevalent in point cloud datasets where both training and test splits follow long-tailed class distributions. To our knowledge, we present the first study on long-tailed point cloud dataset distillation. Rather than focusing primarily on geometric and representation properties or simply constructing a class-balanced synthetic set, our framework explicitly accounts for long-tailed class distributions via two core modules. First, we design Adaptive Synthetic Budgeting to allocate class-wise synthetic budgets according to class quantity and the expected benefit of additional synthetic samples. Given the allocated budgets, we further design 3D Long-Tailed Distribution Matching to optimize synthetic point clouds through Global-Local Feature Alignment and Prior-Aware Supervision. The former preserves both global class distributions and diverse intra-class structures, while the latter provides class-dependent expert supervision to keep tail-class samples recognizable while maintaining diverse head-class patterns. Extensive experiments demonstrate the effectiveness of our method, lifting classification accuracy by 7.0 points on ShapeNet55 against state-of-the-art methods.
