Instance Segmentation of LiDAR Point Clouds

Feihu Zhang, Chenye Guan, Jin Fang, Song Bai, Ruigang Yang, Philip H. S. Torr, Victor PrisacariuICRA 2020 · OpenReview

Abstract

We propose a robust baseline method for instance segmentation which are specially designed for large-scale outdoor LiDAR point clouds. Our method includes a novel dense feature encoding technique, allowing the localization and segmentation of small, far-away objects, a simple but effective solution for single-shot instance prediction and effective strategies for handling severe class imbalances. Since there is no public dataset for the study of LiDAR instance segmentation, we also build a new publicly available LiDAR point cloud dataset to include both precise 3D bounding box and point-wise labels for instance segmentation, while still being about 3~20 times as large as other existing LiDAR datasets. The dataset will be published at https://github.com/feihuzhang/LiDARSeg.
BibTeX
@inproceedings{zhang2020instance,
  author = {Feihu Zhang and Chenye Guan and Jin Fang and Song Bai and Ruigang Yang and Philip H. S. Torr and Victor Prisacariu},
  title = {{Instance Segmentation of LiDAR Point Clouds}},
  booktitle = {IEEE International Conference on Robotics and Automation},
  year = {2020},
  url = {https://openreview.net/forum?id=w8aDYvAkZaTh},
}

Loading related content...

Figure