3D Point Capsule Networks

CVPR 2019 Yongheng ZhaoTolga BirdalHaowen DengFederico Tombari

In this paper, we propose 3D point-capsule networks, an auto-encoder designed to process sparse 3D point clouds while preserving spatial arrangements of the input data. 3D capsule networks arise as a direct consequence of our novel unified 3D auto-encoder formulation... (read more)

PDF Abstract

Evaluation results from the paper

Task Dataset Model Metric name Metric value Global rank Compare
3D Object Classification ModelNet40 3D-PointCapsNet Classification Accuracy 89.3% # 1
3D Part Segmentation ShapeNet-Part 3D-PointCapsNet Accuracy 86% # 1