This repositry is for saving papers related to CV- & geometry-based 6DOF robotic grasping. We MAINLY collect methods utilizing Deep Learning. Note that grasp pose detection related papers are ROUGHLY divided based on their approaches. The list below will be updated in regular basis.
- PoseCNN: A Convolutional Neural Network for 6D Object Pose Estimation in Cluttered Scenes
- SSD-6D: Making RGB-Based 3D Detection and 6D Pose Estimation Great Again
- PVNet: Pixel-wise Voting Network for 6DoF Pose Estimation
- Augmented Autoencoders: Implicit 3D Orientation Learning for 6D Object Detection
- Pix2Pose: Pixel-Wise Coordinate Regression of Objects for 6D Pose Estimation
- DPOD: 6D Pose Object Detector and Refiner
- EPOS: Estimating 6D Pose of Objects with Symmetries
- HybridPose: 6D Object Pose Estimation under Hybrid Representations
- EfficientPose: An efficient, accurate and scalable end-to-end 6D multi object pose estimation approach
- ZebraPose: Coarse to Fine Surface Encoding for 6DoF Object Pose Estimation
- SegICP: Integrated Deep Semantic Segmentation and Pose Estimation
- DenseFusion: 6D Object Pose Estimation by Iterative Dense Fusion
- PVN3D: A Deep Point-wise 3D Keypoints Voting Network for 6DoF Pose Estimation
- MaskUKF: An Instance Segmentation Aided Unscented Kalman Filter for 6D Object Pose and Velocity Tracking
- FFB6D: A Full Flow Bidirectional Fusion Network for 6D Pose Estimation
- Multi-path Learning for Object Pose Estimation Across Domains
- StablePose: Learning 6D Object Poses from Geometrically Stable Patches
- Unseen Object 6D Pose Estimation: A Benchmark and Baselines
- SO(3)-Pose: SO(3)-Equivariance Learning for 6D Object Pose Estimation
Researchers have recently started to investigate methods to make 6D Object Pose Estimation more generalizable. As approaches to enhance generalizability vary, we separated the list from the RGB- and RGBD-based approaches above. A lot of generalizable 6D Object Pose Estimation approaches (#-shot) are multi-stages, and utilize template matching (e.g., support images and CAD models) + classification that NNs are good at.
- EPro-PnP: Generalized End-to-End Probabilistic Perspective-n-Points for Monocular Object Pose Estimation
- OnePose: One-Shot Object Pose Estimation without CAD Models
- Templates for 3D Object Pose Estimation Revisited: Generalization to New Objects and Robustness to Occlusions
- OnePose++: Keypoint-Free One-Shot Object Pose Estimation without CAD Models
- POPE: 6-DoF Promptable Pose Estimation of Any Object, in Any Scene, with One Reference
- Learning to Estimate 6DoF Pose from Limited Data: A Few-Shot, Generalizable Approach using RGB Images
- OVE6D: Object Viewpoint Encoding for Depth-based 6D Object Pose Estimation
- 6-DOF GraspNet: Variational Grasp Generation for Object Manipulation
- UniGrasp: Learning a Unified Model to Grasp with Multifingered Robotic Hands
- 6-DOF Grasping for Target-driven Object Manipulation in Clutter
- Graspness Discovery in Clutters for Fast and Accurate Grasp Detection
- Efficient and Accurate Candidate Generation for Grasp Pose Detection in SE(3)
- UniDexGrasp: Universal Robotic Dexterous Graspingvia Learning Diverse Proposal Generation and Goal-Conditioned Policy
- Data-Efficient Learning for Sim-to-Real Robotic Grasping using Deep Point Cloud Prediction Networks
- Robotic Grasping through Combined Image-Based Grasp Proposal and 3D Reconstruction
- Amodal 3D Reconstruction for Robotic Manipulation via Stability and Connectivity
- Volumetric Grasping Network: Real-time 6 DOF Grasp Detection in Clutter
- GraspNeRF: Multiview-based 6-DoF Grasp Detection for Transparent and Specular Objects Using Generalizable NeRF
- Grasp Pose Detection in Point Clouds
- PointNetGPD: Detecting Grasp Configurations from Point Sets
- Grasp Quality Evaluation Network for Surface-to-Surface Contacts in Point Clouds
- Collision-Aware Target-Driven Object Grasping in Constrained Environments
- GPR: Grasp Pose Refinement Network for Cluttered Scenes
- Simultaneous Semantic and Collision Learning for 6-DoF Grasp Pose Estimation
- DefGraspNets: Grasp Planning on 3D Fields with Graph Neural Nets
- S4G: Amodal Single-view Single-Shot SE(3) Grasp Detection in Cluttered Scenes
- Contact-GraspNet: Efficient 6-DoF Grasp Generation in Cluttered Scenes
- Synergies Between Affordance and Geometry: 6-DoF Grasp Detection via Implicit Representations
- GraspNet-1Billion: A Large-Scale Benchmark for General Object Grasping
- RGB Matters: Learning 7-DoF Grasp Poses on Monocular RGBD Images
- REGNet: REgion-based Grasp Network for End-to-end Grasp Detection in Point Clouds
- Hybrid Physical Metric For 6-DoF Grasp Pose Detection
Researchers have recently started to investigate methods utilizing Neural Radiance Field (NeRF) and Implicit Neural Representation (INR). We separate these approaches as a new category of methods since they more focus on novel representation and sampling for robotic grasping.
- Neural Descriptor Fields: SE(3)-Equivariant Object Representations for Manipulation
- NeuralGrasps: Learning Implicit Representations for Grasps of Multiple Robotic Hands
- Neural Grasp Distance Fields for Robot Manipulation
- Robotic Grasping from Classical to Modern: A Survey
- Deep Learning Approaches to Grasp Synthesis: A Review
- Vision‑based robotic grasping from object localization, object pose estimation to grasp estimation for parallel grippers: a review
| Categorization of approaches in 6D Grasp Pose Detection |
|---|