Here one can find code snippets used in the paper: 3D Object Recognition with Ensemble Learning—A Study of Point Cloud-Based Deep Learning Models springer, arxiv
We focus on the direct point cloud processing because such architectures can be adapted to the analysis of different kinds of sets. Seven architectures are used:
- DeepSets [1]
- PointNet [2]
- PointNet++ [3]
- SO-Net [4]
- KCNet [5]
- DGCNN [6]
- PointCNN [7].
As a part of the project, we set up all above mentioned DL methods on Jetson TX2, which is a reasonable choice for energy efficient mobile robotic applications. We need to set up three different DL frameworks, since PointNet, PointNet++, DGCNN, and PointCNN are using Tensorflow, DeepSets and SO-Net are using PyTorch and KCNet is using Caffe. Here are some tips, how we did it (pls_help).
Jetson TX2 need to be flushed with Jetpack 3.2, this results with CUDA v9.0 and cuDNN v7.0.5. Please do apt update & upgrade afterwards. Here are all dependencies listed:
$ sudo apt-get install python-pip
$ sudo apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev \
libhdf5-dev libhdf5-serial-dev protobuf-compiler
$ sudo apt-get install --no-install-recommends libboost-all-dev
$ sudo apt-get install libgflags-dev libgoogle-glog-dev liblmdb-dev
$ sudo apt-get install libatlas-base-dev libopenblas-devDaniel, PLS, find it on nvidia forum.
First we need to build LLVM from source, I choose the LLVM 7.0.0 version from here.
$ cd llvm-7.0.1.src
$ mkdir build
$ cd build
#Configure to build only the Release version and only for the ARM, x86 and AArch64 architectures
$ cmake .. -DCMAKE_BUILD_TYPE=Release -DLLVM_TARGETS_TO_BUILD="ARM;X86;AArch64"
#Start the build from the build directory
$ cmake -j4 --build .
#install it from the build directory:
$ sudo cmake --build . --target installNext, we can build pytorch:
# clone pyTorch repo
$ git clone http://github.com/pytorch/pytorch
$ cd pytorch
$ git submodule update --init
# install prereqs
$ sudo pip install -U setuptools
$ sudo pip install -r requirements.txt
# Develop Mode:
$ python setup.py build_deps
$ sudo python setup.py developAnd veryfy the installation:
# Verify CUDA (from python interactive terminal)
# import torch
# print(torch.__version__)
# print(torch.cuda.is_available())
# a = torch.cuda.FloatTensor(2)
# print(a)
# b = torch.randn(2).cuda()
# print(b)
# c = a + b
# print(c)KCNet authors prepared cloned version of caffe for KCNet repository. Here's the build procedure:
This work was done with collaboration with Łukasz Chechliński LinkedIn and Tarek El-Gaaly LinkedIn.
[1] M. Zaheer, S. Kottur, S. Ravanbakhsh, B. Poczos, R. R. Salakhutdinov, and A. J. Smola. Deep sets. In Advances in Neural Information Processing Systems, pages 3391–3401, 2017
[2] C. R. Qi, H. Su, K. Mo, and L. J. Guibas. Pointnet: Deep learning on point sets for 3d classification and segmentation. Proc. Computer Vision and Pattern Recognition (CVPR), IEEE, 1(2):4, 2017
[3] C. R. Qi, L. Yi, H. Su, and L. J. Guibas. Pointnet++: Deep hierarchical feature learning on point sets in a metric space. In Advances in Neural Information Processing Systems, pages 5099–5108, 2017
[4] J. Li, B. M. Chen, and G. H. Lee. So-net: Self-organizing network for point cloud analysis. CoRR, abs/1803.04249, 2018
[5] Y. Shen, C. Feng, Y. Yang, and D. Tian. Mining point cloud local structures by kernel correlation and graph pooling. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, volume 4, 2018
[6] Y. Wang, Y. Sun, Z. Liu, S. E. Sarma, M. M. Bronstein, and J. M. Solomon. Dynamic graph cnn for learning on point clouds.arXiv preprint arXiv:1801.07829, 2018
[7] Y. Li, R. Bu, M. Sun, and B. Chen. Pointcnn. arXiv preprint arXiv:1801.07791, 2018