Pretrained backbones with fine-tuning have been widely adopted in 2D vision and natural language processing tasks and demonstrated significant advantages to task-specific networks. In this paper, we present a pretrained 3D backbone, named {\SST}, which first outperforms all state-of-the-art methods in downstream 3D indoor scene understanding tasks. Our backbone network is based on a 3D Swin transformer and carefully designed to efficiently conduct self-attention on sparse voxels with linear memory complexity and capture the irregularity of point signals via generalized contextual relative positional embedding. Based on this backbone design, we pretrained a large {\SST} model on a synthetic Structed3D dataset that is 10 times larger than the ScanNet dataset and fine-tuned the pretrained model in various downstream real-world indoor scene understanding tasks. The results demonstrate that our model pretrained on the synthetic dataset not only exhibits good generality in both downstream segmentation and detection on real 3D point datasets, but also surpasses the state-of-the-art methods on downstream tasks after fine-tuning with +2.3 mIoU and +2.2 mIoU on S3DIS Area5 and 6-fold semantic segmentation, +2.1 mIoU on ScanNet segmentation (val), +1.9 [email protected] on ScanNet detection, +8.1 [email protected] on S3DIS detection. Our method demonstrates the great potential of pretrained 3D backbones with fine-tuning for 3D understanding tasks. The code and models are available at https://github.com/microsoft/Swin3D .

翻译：预训练主干网络结合微调已在二维视觉和自然语言处理任务中得到广泛应用，并展现出相比任务特定网络的显著优势。本文提出一种名为{SST}的预训练3D主干网络，该网络首次在下游3D室内场景理解任务中超越所有最先进方法。我们的主干网络基于3D Swin transformer设计，通过线性内存复杂度在稀疏体素上高效实现自注意力机制，并利用广义上下文相对位置嵌入捕捉点信号的不规则性。基于该主干设计，我们在合成数据集Structed3D（规模是ScanNet数据集的10倍）上预训练了一个大型{SST}模型，并在多种下游真实室内场景理解任务中对预训练模型进行微调。结果表明，在合成数据集上预训练的模型不仅在对真实3D点云数据集的下游分割与检测任务中展现出良好泛化性，而且经微调后在下游任务中超越最先进方法：在S3DIS Area5与6折语义分割上分别提升+2.3 mIoU和+2.2 mIoU，在ScanNet分割（验证集）提升+2.1 mIoU，在ScanNet检测提升+1.9 [email protected]，在S3DIS检测提升+8.1 [email protected]。本方法展示了预训练3D主干网络结合微调在3D理解任务中的巨大潜力。代码与模型已开源至https://github.com/microsoft/Swin3D。