Parameter is Not All You Need: Starting from Non-Parametric Networks for 3D Point Cloud Analysis

14 Mar 2023  ·  Renrui Zhang, Liuhui Wang, Ziyu Guo, Yali Wang, Peng Gao, Hongsheng Li, Jianbo Shi ·

We present a Non-parametric Network for 3D point cloud analysis, Point-NN, which consists of purely non-learnable components: farthest point sampling (FPS), k-nearest neighbors (k-NN), and pooling operations, with trigonometric functions. Surprisingly, it performs well on various 3D tasks, requiring no parameters or training, and even surpasses existing fully trained models. Starting from this basic non-parametric model, we propose two extensions. First, Point-NN can serve as a base architectural framework to construct Parametric Networks by simply inserting linear layers on top. Given the superior non-parametric foundation, the derived Point-PN exhibits a high performance-efficiency trade-off with only a few learnable parameters. Second, Point-NN can be regarded as a plug-and-play module for the already trained 3D models during inference. Point-NN captures the complementary geometric knowledge and enhances existing methods for different 3D benchmarks without re-training. We hope our work may cast a light on the community for understanding 3D point clouds with non-parametric methods. Code is available at https://github.com/ZrrSkywalker/Point-NN.

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
3D Point Cloud Classification ModelNet40 Point-PN Overall Accuracy 93.8 # 22
Number of params 0.8M # 76
Training-free 3D Point Cloud Classification ModelNet40 Point-NN Accuracy (%) 82.6 # 1
Parameters 0M # 1
Need 3D Data? Yes # 1
3D Point Cloud Classification ScanObjectNN Point-PN Overall Accuracy 87.1 # 22
Number of params 0.8M # 39
Training-free 3D Point Cloud Classification ScanObjectNN Point-NN Accuracy (%) 64.9 # 1
Need 3D Data? Yes # 1
Parameters 0M # 1
Training-free 3D Part Segmentation ShapeNet-Part Point-NN mIoU 74.0 # 1
Need 3D Data? Yes # 1
Parameters 0M # 1

Methods