论文标题
稳健点云分类的点 - 素体自适应特征抽象
Point-Voxel Adaptive Feature Abstraction for Robust Point Cloud Classification
论文作者
论文摘要
通过基于学习的方法，已经在云分类中取得了巨大进展。但是，在现实世界应用中，复杂的场景和传感器不准确使点云数据遭受损坏，例如遮挡，噪声和异常值。在这项工作中，我们提出了基于点素的自适应（PV-ADA）特征抽象，以在各种腐败下进行稳健的点云分类。具体而言，提出的框架迭代介绍了点云和提取点素的特征，并具有共享的本地编码和变压器。然后，提出了自适应最大泵来鲁棒地汇总分类的点云特征。 ModelNet-C数据集的实验表明，PV-ADA优于最新方法。特别是，我们将$ 2^{nd} $排名在ModelNet-C分类曲目中，挑战2022，总准确度（OA）为0.865。代码将在https://github.com/zhulf0804/pv-ada上找到。
Great progress has been made in point cloud classification with learning-based methods. However, complex scene and sensor inaccuracy in real-world application make point cloud data suffer from corruptions, such as occlusion, noise and outliers. In this work, we propose Point-Voxel based Adaptive (PV-Ada) feature abstraction for robust point cloud classification under various corruptions. Specifically, the proposed framework iteratively voxelize the point cloud and extract point-voxel feature with shared local encoding and Transformer. Then, adaptive max-pooling is proposed to robustly aggregate the point cloud feature for classification. Experiments on ModelNet-C dataset demonstrate that PV-Ada outperforms the state-of-the-art methods. In particular, we rank the $2^{nd}$ place in ModelNet-C classification track of PointCloud-C Challenge 2022, with Overall Accuracy (OA) being 0.865. Code will be available at https://github.com/zhulf0804/PV-Ada.
