Orientation-boosted Voxel Nets for 3D Object Recognition

Nima Sedaghat, Mohammadreza Zolfaghari, Ehsan Amiri and Thomas Brox

Abstract

Recent work has shown good recognition results in 3D object recognition using 3D convolutional networks. In this paper, we show that the object orientation plays an important role in 3D recognition. More specifically, we argue that objects induce different features in the network under rotation. Thus, we approach the category-level classification task as a multi-task problem, in which the network is trained to predict the pose of the object in addition to the class label as a parallel task. We show that this yields significant improvements in the classification results. We test our suggested architecture on several datasets representing various 3D data sources: LiDAR data, CAD models, and RGB-D images.

Session

Spotlights

Files

Paper (PDF)
Supplementary (PDF)
Video (MP4)

DOI

10.5244/C.31.97
https://dx.doi.org/10.5244/C.31.97

Citation

Bibtex

            @inproceedings{BMVC2017_97,
                title={Orientation-boosted Voxel Nets for 3D Object Recognition},
                author={Nima Sedaghat, Mohammadreza Zolfaghari, Ehsan Amiri and Thomas Brox},
                year={2017},
                month={September},
                pages={97.1-97.13},
                articleno={97},
                numpages={13},
                booktitle={Proceedings of the British Machine Vision Conference (BMVC)},
                publisher={BMVA Press},
                editor={Tae-Kyun Kim, Stefanos Zafeiriou, Gabriel Brostow and Krystian Mikolajczyk},
                doi={10.5244/C.31.97},
                isbn={1-901725-60-X},
                url={https://dx.doi.org/10.5244/C.31.97}
            }