Rotation Invariant Graph Neural Network for 3D Point Clouds

In this paper we propose a novel rotation normalization technique for point cloud processing using an oriented bounding box. We use this method to create a point cloud annotation tool for part segmentation on real camera data. Custom data sets are used to train our network for classification and part segmentation tasks. Successful deployment is completed on an embedded device with limited processing power. A comparison is made with other rotation-invariant features in noisy synthetic datasets. Our method offers more auxiliary information related to the dimension, position, and orientation of the object than previous methods while performing at a similar level.

Automated summary

In this paper, a novel rotation normalization technique using an oriented bounding box for point cloud processing is proposed. It is used to create a point cloud annotation tool for part segmentation and trained on custom datasets for classification and part segmentation tasks. The method is successfully deployed on an embedded device with limited processing power and compared with other rotation-invariant features in noisy synthetic datasets. Our method offers more auxiliary information related to the object's dimension, position, and orientation while performing at a similar level.