TY  - GEN
A1  - Wang, H
A1  - Liu, Q
A1  - Yue, X
A1  - Lasenby, J
A1  - Kusner, MJ
N2  - We describe a simple pre-training approach for point clouds. It works in three steps: 1. Mask all points occluded in a camera view; 2. Learn an encoder-decoder model to reconstruct the occluded points; 3. Use the encoder weights as initialisation for downstream point cloud tasks. We find that even when we pre-train on a single dataset (ModelNet40), this method improves accuracy across different datasets and encoders, on a wide range of downstream tasks. Specifically, we show that our method outperforms previous pre-training methods in object classification, and both part-based and semantic segmentation tasks. We study the pre-trained features and find that they lead to wide downstream minima, have high transformation invariance, and have activations that are highly correlated with part labels. Code and data are available at: https://github.com/hansen7/OcCo.
ID  - discovery10155879
UR  - https://doi.org/10.1109/ICCV48922.2021.00964
PB  - Institute of Electrical and Electronics Engineers (IEEE)
N1  - This version is the author accepted manuscript. For information on re-use, please refer to the publisher's terms and conditions.
TI  - Unsupervised Point Cloud Pre-training via Occlusion Completion
AV  - public
SP  - 9762
Y1  - 2021///
EP  - 9772
ER  -