List of Projects using Minkowski Engine

Please feel free to add your papers on this list.

• A System for General In-Hand Object Re-Orientation, CoRL'21 Best Paper
  • Tags: Dexterous manipulation in robotics, end-to-end training from point cloud to action
• 4D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networks, CVPR'19
  • Tags: Semantic Segmentation, Spatio Temporal Perception
• Spatial Semantic Embedding Network:Fast 3D Instance Segmentation with Deep Metric Learning, arXiv'20
  • Tags: Instance Segmentation, Metric Learning
• Learning multiview 3D point cloud registration, CVPR'20
  • Tags: 3D Registration, Multi-view reconstruction
• Fully Convolutional Geometric Features, ICCV'19
  • Tags: registration, representation learning, metric learning
• PointContrast: Unsupervised Pre-training for 3D Point Cloud Understanding, ECCV'20 spotlight
  • Tags: Unsupervised learning, segmentation, detection, registration
• Deep Global Registration, CVPR'20 oral
  • Tags: 3D reconstruction, registration, high-dimensional convnet
• High-Dimensional Convolutional Networks for Geometric Pattern Recognition, CVPR'20 oral
  • Tags: pattern recognition, 2D registration, 3D registration, high-dimensional convnet
• Generative Sparse Detection Networks for 3D Single-shot Object Detection, ECCV'20 spotlight
  • Tags: Generative Networks, Single-shot detection, 3D YOLO
• Scene-Recognition-in-3D, IROS'20
  • Tags: Scene Recognition, Multi-task
• Sparse Auxiliary Networks for Unified Monocular Depth Prediction and Completion, CVPR'21
  • Tags: Monocular depth prediction, Monocular depth prediction
• Sparse Neighbourhood Consensus Networks, ECCV'20
  • Tags: Image matching, 4D ConvNet
• Deep FusionNet for Point Cloud Semantic Segmentation, ECCV'20
  • Tags: Point Cloud Segmentation
• SparsePipe: Parallel Deep Learning for 3D Point Clouds, HIPC'20
  • Tags: Sparse DNN, 3D point clouds, Distributed Training, Model Parallelism
• Learning to Generate 3D Shapes with Generative Cellular Automata, ICLR'21
  • Tags: Cellular Automata, Generative Model, Shape Generation, Probabilistic Completion
• Hindsight is 20/20: Leveraging Past Traversals to Aid 3D Perception, ICLR'22
  • Tags: 3D detection with historical context