Identifying Unknown Instances for Autonomous Driving

Kelvin Wong, Shenlong Wang, Mengye Ren, Ming Liang, Raquel Urtasun
Proceedings of the Conference on Robot Learning, PMLR 100:384-393, 2020.

Abstract

In the past few years, we have seen great progress in perception algorithms, particular through the use of deep learning. However, most existing approaches focus on a few categories of interest, which represent only a small fraction of the potential categories that robots need to handle in the real-world. Thus, identifying objects from unknown classes remains a challenging yet crucial task. In this paper, we develop a novel open-set instance segmentation algorithm for point clouds which can segment objects from both known and unknown classes in a holistic way. Our method uses a deep convolutional neural network to project points into a category-agnostic embedding space in which they can be clustered into instances irrespective of their semantics. Experiments on two large-scale self-driving datasets validate the effectiveness of our proposed method.

Cite this Paper


BibTeX
@InProceedings{pmlr-v100-wong20a, title = {Identifying Unknown Instances for Autonomous Driving}, author = {Wong, Kelvin and Wang, Shenlong and Ren, Mengye and Liang, Ming and Urtasun, Raquel}, booktitle = {Proceedings of the Conference on Robot Learning}, pages = {384--393}, year = {2020}, editor = {Kaelbling, Leslie Pack and Kragic, Danica and Sugiura, Komei}, volume = {100}, series = {Proceedings of Machine Learning Research}, month = {30 Oct--01 Nov}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v100/wong20a/wong20a.pdf}, url = {https://proceedings.mlr.press/v100/wong20a.html}, abstract = {In the past few years, we have seen great progress in perception algorithms, particular through the use of deep learning. However, most existing approaches focus on a few categories of interest, which represent only a small fraction of the potential categories that robots need to handle in the real-world. Thus, identifying objects from unknown classes remains a challenging yet crucial task. In this paper, we develop a novel open-set instance segmentation algorithm for point clouds which can segment objects from both known and unknown classes in a holistic way. Our method uses a deep convolutional neural network to project points into a category-agnostic embedding space in which they can be clustered into instances irrespective of their semantics. Experiments on two large-scale self-driving datasets validate the effectiveness of our proposed method.} }
Endnote
%0 Conference Paper %T Identifying Unknown Instances for Autonomous Driving %A Kelvin Wong %A Shenlong Wang %A Mengye Ren %A Ming Liang %A Raquel Urtasun %B Proceedings of the Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2020 %E Leslie Pack Kaelbling %E Danica Kragic %E Komei Sugiura %F pmlr-v100-wong20a %I PMLR %P 384--393 %U https://proceedings.mlr.press/v100/wong20a.html %V 100 %X In the past few years, we have seen great progress in perception algorithms, particular through the use of deep learning. However, most existing approaches focus on a few categories of interest, which represent only a small fraction of the potential categories that robots need to handle in the real-world. Thus, identifying objects from unknown classes remains a challenging yet crucial task. In this paper, we develop a novel open-set instance segmentation algorithm for point clouds which can segment objects from both known and unknown classes in a holistic way. Our method uses a deep convolutional neural network to project points into a category-agnostic embedding space in which they can be clustered into instances irrespective of their semantics. Experiments on two large-scale self-driving datasets validate the effectiveness of our proposed method.
APA
Wong, K., Wang, S., Ren, M., Liang, M. & Urtasun, R.. (2020). Identifying Unknown Instances for Autonomous Driving. Proceedings of the Conference on Robot Learning, in Proceedings of Machine Learning Research 100:384-393 Available from https://proceedings.mlr.press/v100/wong20a.html.

Related Material