3D Deep Learning works
- Taco S. Cohen, Spherical CNNs, ICLR 2018 Best paper [paper]
- Learning SO(3) Equivariant Representations with Spherical CNNs [paper] [code]
- 3D CNN
- 3D-DenseNet
- Voxnet: A 3d convolutional neural network for real-time object recognition, IROS 2015. [code] [paper]
- [3D-NIN, network in network]
- VRN Ensemble, Generative and discriminative voxel modeling with convolutional neural networks, arxiv [paper] [code]
- Voxception-Resnet Blocks
- 2D CNN - MVCNN, Learned-Miller.Multi- view convolutional neural networks for 3d shape recognition, ICCV2015 [project] [code] [paper][data] [video]
- Point
- Graph/tree-based
- Kd-Net, scape from cells: Deep kd- networks for the recognition of 3d point cloud models, arxiv2017 [paper]
- kd-tree
- Octnet: Learning deep 3d representations at high resolutions, CVPR2017
- octree
- O-cnn: Octree-based convolutional neural networks for 3d shape analysis, TOG2017
- octree
- SO-Net, SO-Net: Self-Organizing Network for Point Cloud Analysis, CVPR2018 [paper] [paper] [code]
- point-to-node kNN search Self-Organizing Map (SOM)
- KCNet, Mining Point Cloud Local Structures by Kernel Correlation and Graph Pooling, CVPR2018 [project] [code] [paper][data] [video]
- Kernel Correlation
- Graph Pooling
- Kd-Net, scape from cells: Deep kd- networks for the recognition of 3d point cloud models, arxiv2017 [paper]
Data types: RGBD, Flow, Laser
- MV3D, Multi-View 3D Object Detection Network for Autonomous Driving [paper] [code]
- Avod, Joint 3D Proposal Generation and Object Detection from View Aggregation [paper] [code]
- F-PointNet, Frustum PointNets for 3D Object Detection from RGB-D Data [paper] [code]
- VoxelNet: End-to-End Learning for Point Cloud Based 3D Object Detection [paper]
Data types: RGBD, Flow, Laser
- ShapeNet
- SO-Net
- 3D-GAN
- CVPR2016 Tutorial: 3D Deep Learning with Marvin
- 3D Shape Retrieval
- C3D, website
- [Video Caffe(C3D)] [code]
- [DeepMedic, Brain Lesion Segmentation] [[code(https://github.com/Kamnitsask/deepmedic)]
- 3D Keypoint Detection and Feature Description
- Learning 3D Object Orientations From Synthetic Images
- Read
LSTM: A Search Space Odyssey
and implement LSTM. - Use Tree LSTM in place of LSTM.
Paper
Code
Slides
- Learning 3D Object Orientations From Synthetic Images
- 3D Shape Segmentation with Projective Convolutional Networks. CVPR2017.
Project
Poster
Presentation
-
more usefull tools should be added in.
- Read and process process data as needed by the network.
- Read
Learning Phrase Representations using RNN Encoder–Decoder for Statistical Machine Translation
and implement a GRU.