VoxelNet: End-to-End Learning for Point Cloud Based 3D Object Detection

@inproceedings{Zhou2018VoxelNetEL,
  title={VoxelNet: End-to-End Learning for Point Cloud Based 3D Object Detection},
  author={Yin Zhou and Oncel Tuzel},
  booktitle={CVPR},
  year={2018}
}
The main purpose of subsampling points, for the voxels that have more than a predefined maximum number of points, is to form an efficient tensor representation for computation on a GPU. In our experiments, the maximum number of points is set to a large enough value, T = 35 for car detection and 45 for pedestrian/cyclist detection, such that the aggregated statistics within the voxels, e.g. mean or other pooling operations, are close to the true statistics. On the datasets presented, only 0.17… CONTINUE READING
Highly Influential
This paper has highly influenced 13 other papers. REVIEW HIGHLY INFLUENTIAL CITATIONS
Highly Cited
This paper has 78 citations. REVIEW CITATIONS

Citations

Publications citing this paper.
Showing 1-10 of 47 extracted citations

78 Citations

05010020172018
Citations per Year
Semantic Scholar estimates that this publication has 78 citations based on the available data.

See our FAQ for additional information.

References

Publications referenced by this paper.
Showing 1-10 of 48 references

Onboard object detection: Multicue, multimodal, and multiview random forest of local experts

  • A. Gonzlez, D. Vzquez, A. M. Lpez, J. Amores
  • IEEE Transactions on Cybernetics,
  • 2017
1 Excerpt

Similar Papers

Loading similar papers…