Revisiting Point Cloud Shape Classification with a Simple and Effective Baseline

By Ankit Goyal, Hei Law, Bowei Liu, Alejandro Newell, Jia Deng. International Conference on Machine Learning, 2021

308 citations recorded by the source corpus.

Processing point cloud data is an important component of many real-world systems. As such, a wide variety of point-based approaches have been proposed, reporting steady benchmark improvements over time. We study the key ingredients of this progress and uncover two critical results. First, we find that auxiliary factors like different evaluation schemes, data augmentation strategies, and loss functions, which are independent of the model architecture, make a large difference in performance. The differences are large enough that they obscure the effect of architecture. When these factors are controlled for, PointNet++, a relatively older network, performs competitively with recent methods. Second, a very simple projection-based method, which we refer to as SimpleView, performs surprisingly well. It achieves on par or better results than sophisticated state-of-the-art methods on ModelNet40 while being half the size of PointNet++. It also outperforms state-of-the-art methods on ScanObjectNN, a real-world point cloud benchmark, and demonstrates better cross-dataset generalization. Code is available at https://github.com/princeton-vl/SimpleView.

Subfields: Computational Mechanics, Computer Graphics and Computer-Aided Design, Geology

View Paper: https://www.semanticscholar.org/paper/4e5dff8d9db8ee21b636e9bf93a2fc96999079aa

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

1. Yu Qiao
2. Bin Cui
3. Peng Gao
4. Dieter Fox
5. Kunchang Li
6. Renrui Zhang
7. Hongsheng Li
8. Ankit Goyal

In Computer Graphics and Computer-Aided Design

1. Yu Qiao
2. Dieter Fox
3. Hongsheng Li

In Geology

1. Yu Qiao
2. Peng Gao
3. Dieter Fox
4. Renrui Zhang
5. Hongsheng Li
6. Ankit Goyal
7. Ziyu Guo

Citations in the corpus

• DiffusionNet: Discretization Agnostic Learning on Surfaces (2020, unknown): “2 (see discussion in [Goyal et al. 2021]).”
• PointCLIP: Point Cloud Understanding by CLIP (2021, methodology): “As is presented in Table 6, ensembling PointNet++ lowers the performance of RSCNN and CurveNet, and aggregating the highest two models, SimpleView and CurveNet, could not achieve better performance.”
• RVT: Robotic View Transformer for 3D Object Manipulation (2023, methodology): “For point cloud recognition, SimpleView [52] showed how a simple view-based method outperforms sophisticated point-based methods.”
• PointNeXt: Revisiting PointNet++ with Improved Training and Scaling Strategies (2022, unknown): “In the point cloud domain, SimpleView [10] is the first work to show that training strategies have a large impact on the performance of a neural network.”
• Rethinking Network Design and Local Geometry in Point Cloud: A Simple Residual MLP Framework (2022, unknown): “(Xu et al., 2018) 69.8 73.7 PointNet++ (Qi et al., 2017b) 75.4 77.9 DGCNN (Wang et al., 2019) 73.6 78.1 PointCNN (Li et al., 2018b) 75.1 78.5 BGA-DGCNN (Uy et al., 2019) 75.7 79.7 BGA-PN++ (Uy et al., 2019) 77.5 80.2 DRNet (Qiu et al., 2021a) 78.0 80.3 GBNet (Qiu et al., 2021b) 77.8 80.5 SimpleView ”