LinNet: Linear Network for Efficient Point Cloud Representation Learning

Point-based methods have made significant progress, but improving their scalability in large-scale 3D scenes is still a challenging problem. In this paper, we delve into the point-based method and develop a simpler, faster, stronger variant model, dubbed as LinNet. In particular, we first propose the disassembled set abstraction (DSA) module, which is more effective than the previous version of set abstraction. It achieves more efficient local aggregation by leveraging spatial anisotropy and channel anisotropy separately. Additionally, by mapping 3D point clouds onto 1D space-filling curves, we enable parallelization of downsampling and neighborhood queries on GPUs with linear complexity. LinNet, as a purely point-based method, outperforms most previous methods in both indoor and outdoor scenes without any extra attention, and sparse convolution but merely relying on a simple MLP. It achieves the mIoU of 73. 7\%, 81. 4\%, and 69. 1\% on the S3DIS Area5, NuScenes, and SemanticKITTI validation benchmarks, respectively, while speeding up almost 10x times over PointNeXt. Our work further reveals both the efficacy and efficiency potential of the vanilla point-based models in large-scale representation learning. Our code will be available upon publication.

Authors

• Hao Deng Huazhong Agricultural University Engineering Research Center of Intelligent Technology for Agriculture, Ministry of Education
• Kunlei Jing
• Shengmei Cheng
• Cheng Liu Chongqing Key Laboratory of Computational Intelligence Chongqing University of Posts and Telecommunications, Chongqing, China Key Laboratory of Cyberspace Big Data Intelligent Security, Ministry of Education
• Jiawei Ru
• Jiang Bo
• Lin Wang

Keywords

No keywords are indexed for this paper.