Here you can find the code for the BMVC 2017 version of "Orientation-boosted Voxel Nets for 3D Object Recognition", a.k.a ORION.
If you use this code for research please cite:
@InProceedings{SZB17a,
author = "N. Sedaghat and M. Zolfaghari and E. Amiri and T. Brox",
title = "Orientation-boosted voxel nets for 3D object recognition",
booktitle = "British Machine Vision Conference (BMVC)",
month = " ",
year = "2017",
url = "http://lmb.informatik.uni-freiburg.de/Publications/2017/SZB17a"
}
Please download, build and install the version of caffe introduced here -- we use it for the provided 3D augmentation capabilities.
Modelnet40 aligned objects can be downloaded from the links below. If you use this modified dataset, please also cite the following work from which we borrowed techniques for the auto-alignment:
@InProceedings{SB15,
author = "N. Sedaghat and T. Brox",
title = "Unsupervised Generation of a Viewpoint Annotated Car Dataset from Videos",
booktitle = "IEEE International Conference on Computer Vision (ICCV)",
year = "2015",
url = "http://lmb.informatik.uni-freiburg.de/Publications/2015/SB15"
}
Auto-alignment code will be published soon.