arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2207.00531cs.CVcs.LG
收藏

用于激光雷达点云自监督预训练的掩码自编码器

Masked Autoencoder for Self-Supervised Pre-training on Lidar Point Clouds

Chalmers University of Technology(查尔姆斯理工大学);Zenseact

Georg Hess, Johan Jaxing, Elias Svensson, David Hagerman, Christoffer Petersson, Lennart Svensson

摘要:

Masked autocoding已经成为文本、图像和最近的点云的Transformer模型的成功预训练范例。原始汽车数据集是自监督预训练的合适候选者,因为与3D目标检测(OD)等任务的注释相比,它们的收集成本通常较低。然而,用于点云的掩码自编码器的开发仅集中在合成和室内数据上。因此,现有的方法已经针对具有均匀点密度的小而密集的点云定制了它们的表示和模型。在这项工作中,我们研究了掩蔽自动编码的点云在汽车设置,这是稀疏的,其中的点密度可以在同一场景中的对象之间的显着变化。为此,我们提出了体素-MAE,一个简单的掩蔽自动编码预训练方案,专为体素表示。我们预先训练了基于Transformer的3D物体检测器的主干,以重建被掩蔽的体素,并区分空体素和非空体素。我们的方法在具有挑战性的nuScenes数据集上将3D OD性能提高了1.75 mAP点和1.05 NDS。此外,我们表明,通过使用Voxel-MAE进行预训练,我们只需要40%的注释数据就可以超越随机初始化的等价数据。代码可在https://github.com/georghess/voxel-mae获得

英文摘要:

Masked autoencoding has become a successful pretraining paradigm for Transformer models for text, images, and, recently, point clouds. Raw automotive datasets are suitable candidates for self-supervised pre-training as they generally are cheap to collect compared to annotations for tasks like 3D object detection (OD). However, the development of masked autoencoders for point clouds has focused solely on synthetic and indoor data. Consequently, existing methods have tailored their representations and models toward small and dense point clouds with homogeneous point densities. In this work, we study masked autoencoding for point clouds in an automotive setting, which are sparse and for which the point density can vary drastically among objects in the same scene. To this end, we propose Voxel-MAE, a simple masked autoencoding pre-training scheme designed for voxel representations. We pre-train the backbone of a Transformer-based 3D object detector to reconstruct masked voxels and to distinguish between empty and non-empty voxels. Our method improves the 3D OD performance by 1.75 mAP points and 1.05 NDS on the challenging nuScenes dataset. Further, we show that by pre-training with Voxel-MAE, we require only 40% of the annotated data to outperform a randomly initialized equivalent. Code available at https://github.com/georghess/voxel-mae

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑