arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2312.10035cs.CV
收藏

Point Transformer V3:更简单、更快速、更强

Point Transformer V3: Simpler, Faster, Stronger

HKU SH AI Lab(香港大学深圳人工智能实验室);CUHK(SZ)(香港中文大学(深圳));PKU(北京大学);MIT(麻省理工学院)

Xiaoyang Wu, Li Jiang, Peng-Shuai Wang, Zhijian Liu, Xihui Liu, Yu Qiao, Wanli Ouyang, Tong He, Hengshuang Zhao

AI总结:

提出Point Transformer V3,通过序列化邻居映射替代KNN实现简单高效的点云处理,在20多个任务上取得SOTA结果。

摘要:

本文不以在注意力机制内寻求创新为动机。相反,它专注于克服点云处理背景下准确性与效率之间现有的权衡,利用规模的力量。受3D大规模表示学习最新进展的启发,我们认识到模型性能更多受规模影响,而非复杂设计。因此,我们提出Point Transformer V3(PTv3),它优先考虑简单性和效率,而非某些机制在扩展后对整体性能影响较小的准确性,例如用基于特定模式组织的点云的高效序列化邻居映射替代KNN的精确邻居搜索。这一原则实现了显著扩展,将感受野从16个点扩展到1024个点,同时保持高效(与前代PTv2相比,处理速度提升3倍,内存效率提升10倍)。PTv3在跨越室内外场景的20多个下游任务上取得最先进结果。通过多数据集联合训练进一步增强,PTv3将这些结果推向更高水平。

英文摘要:

This paper is not motivated to seek innovation within the attention mechanism. Instead, it focuses on overcoming the existing trade-offs between accuracy and efficiency within the context of point cloud processing, leveraging the power of scale. Drawing inspiration from recent advances in 3D large-scale representation learning, we recognize that model performance is more influenced by scale than by intricate design. Therefore, we present Point Transformer V3 (PTv3), which prioritizes simplicity and efficiency over the accuracy of certain mechanisms that are minor to the overall performance after scaling, such as replacing the precise neighbor search by KNN with an efficient serialized neighbor mapping of point clouds organized with specific patterns. This principle enables significant scaling, expanding the receptive field from 16 to 1024 points while remaining efficient (a 3x increase in processing speed and a 10x improvement in memory efficiency compared with its predecessor, PTv2). PTv3 attains state-of-the-art results on over 20 downstream tasks that span both indoor and outdoor scenarios. Further enhanced with multi-dataset joint training, PTv3 pushes these results to a higher level.

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。

↑