TY - GEN
T1 - Point Cloud Augmentation with Weighted Local Transformations
AU - Kim, Sihyeon
AU - Lee, Sanghyeok
AU - Hwang, Dasol
AU - Lee, Jaewon
AU - Hwang, Seong Jae
AU - Kim, Hyunwoo J.
N1 - Funding Information:
Acknowledgments. This work was supported by ICT Creative Consilience program(IITP-2021-2020-0-01819) supervised by the IITP, Research on CPU vulnerability detection and validation (No. 2019-0-00533), the National Research Council of Science & Technology (NST) grant by the Korea government (MSIT) (No. CAP-18-03-ETRI), and Samsung Electronics.
Publisher Copyright:
© 2021 IEEE
PY - 2021
Y1 - 2021
N2 - Despite the extensive usage of point clouds in 3D vision, relatively limited data are available for training deep neural networks. Although data augmentation is a standard approach to compensate for the scarcity of data, it has been less explored in the point cloud literature. In this paper, we propose a simple and effective augmentation method called PointWOLF for point cloud augmentation. The proposed method produces smoothly varying non-rigid deformations by locally weighted transformations centered at multiple anchor points. The smooth deformations allow diverse and realistic augmentations. Furthermore, in order to minimize the manual efforts to search the optimal hyperparameters for augmentation, we present AugTune, which generates augmented samples of desired difficulties producing targeted confidence scores. Our experiments show our framework consistently improves the performance for both shape classification and part segmentation tasks. Particularly, with PointNet++, PointWOLF achieves the state-of-the-art 89.7 accuracy on shape classification with the real-world ScanObjectNN dataset. The code is available at https://github.com/mlvlab/PointWOLF.
AB - Despite the extensive usage of point clouds in 3D vision, relatively limited data are available for training deep neural networks. Although data augmentation is a standard approach to compensate for the scarcity of data, it has been less explored in the point cloud literature. In this paper, we propose a simple and effective augmentation method called PointWOLF for point cloud augmentation. The proposed method produces smoothly varying non-rigid deformations by locally weighted transformations centered at multiple anchor points. The smooth deformations allow diverse and realistic augmentations. Furthermore, in order to minimize the manual efforts to search the optimal hyperparameters for augmentation, we present AugTune, which generates augmented samples of desired difficulties producing targeted confidence scores. Our experiments show our framework consistently improves the performance for both shape classification and part segmentation tasks. Particularly, with PointNet++, PointWOLF achieves the state-of-the-art 89.7 accuracy on shape classification with the real-world ScanObjectNN dataset. The code is available at https://github.com/mlvlab/PointWOLF.
UR - http://www.scopus.com/inward/record.url?scp=85124639862&partnerID=8YFLogxK
U2 - 10.1109/ICCV48922.2021.00059
DO - 10.1109/ICCV48922.2021.00059
M3 - Conference contribution
AN - SCOPUS:85124639862
T3 - Proceedings of the IEEE International Conference on Computer Vision
SP - 528
EP - 537
BT - Proceedings - 2021 IEEE/CVF International Conference on Computer Vision, ICCV 2021
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 18th IEEE/CVF International Conference on Computer Vision, ICCV 2021
Y2 - 11 October 2021 through 17 October 2021
ER -