作者：Tom Hardy
點擊上方“3D視覺工坊”，選擇“星標”
干貨第一時間送達
作者丨Tom Hardy@知乎
來源丨https://zhuanlan.zhihu.com/p/305073693
編輯丨3D視覺工坊
姿態估計算法匯總|基於RGB、RGB-D以及點雲數據
主要有整體方式、霍夫投票方式、Keypoint-based方式、Dense Correspondence方式等。
實現方法：傳統方法、深度學習方式。
數據不同：RGB、RGB-D、點雲數據等；
標注工具
LabelFusion：https://github.com/RobotLocomotion/LabelFusion
實現方式不同
整體方式
整體方法直接估計給定圖像中物體的三維位置和方向。經典的基於模板的方法構造剛性模板並掃描圖像以計算最佳匹配姿態。這種手工制作的模板對集群場景不太可靠。最近，人們提出了一些基於深度神經網絡的方法來直接回歸相機或物體的6D姿態。然而，旋轉空間的非線性使得數據驅動的DNN難以學習和推廣。
1.Discriminative mixture-of-templates for viewpoint classification
2.Gradient response maps for realtime detection of textureless objects.
3.Comparing images using the hausdorff distance
4.Implicit 3d orientation learning for 6d object detection from rgb images.
5.Instance- and Category-level 6D Object Pose Estimation
基於模型
1.Matching RGB Images to CAD Models for Object Pose Estimation - Georgios Georgakis, Srikrishna Karanam, Ziyan Wu, and Jana Kosecka. [Paper：https://arxiv.org/pdf/1811.07249.pdf]
2.Deep model-based 6d pose refinement in rgb
Keypoint-based方式
目前基於關鍵點的方法首先檢測圖像中物體的二維關鍵點，然后利用PnP算法估計6D姿態。
1.Surf: Speeded up robust features.
2.Object recognition from local scaleinvariant features
3.3d object modeling and recognition using local affine-invariant image descriptors and multi-view spatial constraints.
4.DeepIM: Deep Iterative Matching for 6D Pose Estimation - Yi Li, Gu Wang, Xiangyang Ji, Yu Xiang, Dieter Fox. [Paper：https://arxiv.org/pdf/1804.00175.pdf]
5.Stacked hourglass networks for human pose estimation
6.Making deep heatmaps robust to partial occlusions for 3d object pose estimation.
7.Bb8: A scalable, accurate, robust to partial occlusion method for predicting the 3d poses of challenging objects without using depth
8.Real-time seamless single shot 6d object pose prediction.
9.Discovery of latent 3d keypoints via end-toend geometric reasoning.
10.Pvnet: Pixel-wise voting network for 6dof pose estimation.
11.Single-Stage 6D Object Pose Estimation - Yinlin Hu,Pascal Fua,Wei Wang,Mathieu Salzmann. [Paper：https://arxiv.org/pdf/1911.08324.pdf]
12.Estimating 6D Pose From Localizing Designated Surface Keypoints - Zelin Zhao, Gao Peng, Haoyu Wang, Hao-Shu Fang, Chengkun Li, Cewu Lu. [Paper：https://arxiv.org/pdf/1812.01387v1.pdf]
13.Learning 6D Object Pose Estimation Using 3D Object Coordinates - Eric Brachmann, Alexander Krull, Frank Michel, Stefan Gumhold, Jamie Shotton, Carsten Rother. [Paper：https://link.springer.com/content/pdf/10.1007%2F978-3-319-10605-2_35.pdf]
Dense Correspondence/霍夫投票方式
1.Independent object class detection using 3d feature maps.
2.Depth encoded hough voting for joint object detection and shape recovery.
3.aware object detection and pose estimation.
4.Learning 6d object pose estimation using 3d object coordinates.
5.Global hypothesis generation for 6d object pose estimation.
6.Deep learning of local rgb-d patches for 3d object detection and 6d pose estimation.
7.Cdpn: Coordinates-based disentangled pose network for real-time rgb-based 6-dof object pose estimation.
8.Pix2pose: Pixel-wise coordinate regression of objects for 6d pose estimation.
9.Normalized object coordinate space for categorylevel 6d object pose and size estimation.
10.Recovering 6d object pose and predicting next-bestview in the crowd.
11.PVNet: Pixel-wise Voting Network for 6DoF Pose Estimation - Sida Peng, Yuan Liu, Qixing Huang, Xiaowei Zhou, Hujun Bao. [Paper：https://arxiv.org/pdf/1812.11788.pdf]
基於分割
1.Segmentation-driven 6D Object Pose Estimation - Yinlin Hu, Joachim Hugonot, Pascal Fua, Mathieu Salzmann. [Paper：https://arxiv.org/pdf/1812.02541.pdf]
2.Deep Object Pose Estimation for Semantic Robotic Grasping of Household Objects - Jonathan Tremblay, Thang To, Balakumar Sundaralingam, Yu Xiang, Dieter Fox, Stan Birchfield. [Paper：https://arxiv.org/pdf/1809.10790.pdf]
深度學習相關方法
1.PoseCNN: A convolutional neural network for 6d object pose estimation in cluttered scenes.
2.Render for cnn: Viewpoint estimation in images using cnns trained with rendered 3d model views.
3.Segmentation-driven 6D Object Pose Estimation - Yinlin Hu, Joachim Hugonot, Pascal Fua, Mathieu Salzmann.[Paper：https://openaccess.thecvf.com/content_CVPR_2019/papers/Hu_Segmentation-Driven_6D_Object_Pose_Estimation_CVPR_2019_paper.pdf]
4.Single-Stage 6D Object Pose Estimation - Yinlin Hu,Pascal Fua,Wei Wang,Mathieu Salzmann. [Paper：https://arxiv.org/pdf/1911.08324.pdf]
5.Normalized Object Coordinate Space for Category-Level 6D Object Pose and Size Estimation - He Wang, Srinath Sridhar, Jingwei Huang, Julien Valentin, Shuran Song, Leonidas J. Guibas. [Paper：https://arxiv.org/pdf/1901.02970v1.pdf]
6.Robust 6D Object Pose Estimation in Cluttered Scenesusing Semantic Segmentation and Pose Regression Networks - Arul Selvam Periyasamy, Max Schwarz, and Sven Behnke. [[Paper]
7.Implicit 3D Orientation Learning for 6D Object Detection from RGB Images：https://www.ais.uni-bonn.de/papers/IROS_2018_Periyasamy.pdf
8.DenseFusion: 6D Object Pose Estimation by Iterative Dense Fusion - Chen Wang, Danfei Xu, Yuke Zhu, Roberto Martín-Martín, Cewu Lu, Li Fei-Fei, Silvio Savarese. [Paper：https://arxiv.org/pdf/1901.04780.pdf]
9.Real-Time Object Pose Estimation with Pose Interpreter Networks- Jimmy Wu, Bolei Zhou, Rebecca Russell, Vincent Kee, Syler Wagner, Mitchell Hebert, Antonio Torralba, David M.S. Johnson. [Paper：https://arxiv.org/pdf/1808.01099.pdf]
10.BB8: A Scalable, Accurate, Robust to Partial Occlusion Method for Predicting the 3D Poses of Challenging Objects without Using Depth - Mahdi Rad, Vincent Lepetit. [Paper：https://arxiv.org/abs/1703.10896]
11.Real-Time Seamless Single Shot 6D Object Pose Prediction - Bugra Tekin, Sudipta N. Sinha, Pascal Fua. [Paper：https://arxiv.org/pdf/1711.08848.pdf]
12.SSD-6D: Making RGB-based 3D detection and 6D pose estimation great again - Wadim Kehl, Fabian Manhardt, Federico Tombari, Slobodan Ilic, Nassir Navab. [Paper：https://arxiv.org/pdf/1711.10006.pdf]
13.Deep Learning of Local RGB-D Patches for 3D Object Detection and 6D Pose Estimation - Wadim Kehl, Fausto Milletari, Federico Tombari, Slobodan Ilic, Nassir Navab. [Paper：https://arxiv.org/pdf/1607.06038.pdf]
14.Deep-6DPose: Recovering 6D Object Pose from a Single RGB Image - Thanh-Toan Do, Ming Cai, Trung Pham, Ian Reid：https://arxiv.org/pdf/1802.10367.pdf
數據格式不同
根據數據格式的不同，又可分為基於RGB、RGB-D、點雲數據的識別算法。
基於點雲方式
1.PointFusion
2.Frustum PointNets
3.VoteNet
基於RGB方式
1.SilhoNet: An RGB Method for 6D Object Pose Estimation - Gideon Billings, Matthew Johnson-Roberson. [Paper：https://arxiv.org/pdf/1809.06893.pdf]
2.PVNet: Pixel-wise Voting Network for 6DoF Pose Estimation - Sida Peng, Yuan Liu, Qixing Huang, Xiaowei Zhou, Hujun Bao. [Paper：https://arxiv.org/pdf/1812.11788.pdf]
3.DPOD: 6D Pose Object Detector and Refiner - Sergey Zakharov, Ivan Shugurov, Slobodan Ilic. [Paper：https://arxiv.org/pdf/1902.11020v2.pdf]
4.Deep model-based 6d pose refinement in rgb
5.Implicit 3D Orientation Learning for 6D Object Detection from RGB Images - Martin Sundermeyer, Zoltan-Csaba Marton, Maxmilian Durner, Manuel Brucker and Rudolph Triebel. [Paper：https://arxiv.org/pdf/1902.01275v1.pdf]
6.BB8: A Scalable, Accurate, Robust to Partial Occlusion Method for Predicting the 3D Poses of Challenging Objects without Using Depth - Mahdi Rad, Vincent Lepetit. [Paper：https://arxiv.org/abs/1703.10896]
7.Real-Time Seamless Single Shot 6D Object Pose Prediction - Bugra Tekin, Sudipta N. Sinha, Pascal Fua. [Paper：https://arxiv.org/pdf/1711.08848.pdf]
8.SSD-6D: Making RGB-based 3D detection and 6D pose estimation great again - Wadim Kehl, Fabian Manhardt, Federico Tombari, Slobodan Ilic, Nassir Navab. [Paper：https://arxiv.org/pdf/1711.10006.pdf]
9.Deep-6DPose: Recovering 6D Object Pose from a Single RGB Image - Thanh-Toan Do, Ming Cai, Trung Pham, Ian Reid:https://arxiv.org/pdf/1802.10367.pdf
基於RGB-D方式
1.Category-level 6D Object Pose Recovery in Depth Images - Caner Sahin and Tae-Kyun Kim. [Paper:https://openaccess.thecvf.com/content_ECCVW_2018/papers/11129/Sahin_Category-level_6D_Object_Pose_Recovery_in_Depth_Images_ECCVW_2018_paper.pdf]
2.Holistic and local patch framework for 6D object pose estimation in RGB-D images - Haoruo Zhang, Qixin Cao. [Paper:https://www.sciencedirect.com/science/article/abs/pii/S1077314219300050]
3.Multi-view 6D Object Pose Estimation and Camera Motion Planning Using RGBD Images - Juil Sock, S. Hamidreza Kasaei, Luís Seabra Lopes, Tae-Kyun Kim. [Paper:https://ieeexplore.ieee.org/document/8265470]
4.Deep Learning of Local RGB-D Patches for 3D Object Detection and 6D Pose Estimation - Wadim Kehl, Fausto Milletari, Federico Tombari, Slobodan Ilic, Nassir Navab. [Paper:https://arxiv.org/pdf/1607.06038.pdf]
相關源碼
1.HybridPose：https://github.com/chensong1995/HybridPose
2.PoseCNN：https://github.com/yuxng/PoseCNN
3.Single Shot Pose Estimation:https://github.com/Microsoft/singleshotpose
SSD-6D：https://link.zhihu.com/?target=https%3A//github.com/wadimkehl/ssd-6d
4.Dope Object Pose Estimation：https://github.com/NVlabs/Deep_Object_Pose
5.Pose Interpreter Networks：https://github.com/jimmyyhwu/pose-interpreter-networks
6.Tools for Evaluation of 6D Object Pose Estimation:https://github.com/thodan/obj_pose_eval
7.Augmented Autoencoder:https://github.com/DLR-RM/AugmentedAutoencoder
8.DeepIM：https://github.com/liyi14/mx-DeepIM
9.DenseFusion：https://github.com/j96w/DenseFusion
10.BetaPose：https://github.com/sjtuytc/betapose
11.PVNet:https://link.zhihu.com/?12.target=https%3A//github.com/zju3dv/pvnet
本文僅做學術分享，如有侵權，請聯系刪文。
下載1
在「3D視覺工坊」公眾號后台回復：3D視覺，即可下載 3D視覺相關資料干貨，涉及相機標定、三維重建、立體視覺、SLAM、深度學習、點雲后處理、多視圖幾何等方向。
下載2
在「3D視覺工坊」公眾號后台回復：3D視覺github資源匯總，即可下載包括結構光、標定源碼、缺陷檢測源碼、深度估計與深度補全源碼、點雲處理相關源碼、立體匹配源碼、單目、雙目3D檢測、基於點雲的3D檢測、6D姿態估計源碼匯總等。
下載3
在「3D視覺工坊」公眾號后台回復：相機標定，即可下載獨家相機標定學習課件與視頻網址；后台回復：立體匹配，即可下載獨家立體匹配學習課件與視頻網址。
重磅！3DCVer-學術論文寫作投稿 交流群已成立掃碼添加小助手微信，可申請加入3D視覺工坊-學術論文寫作與投稿 微信交流群，旨在交流頂會、頂刊、SCI、EI等寫作與投稿事宜。
同時也可申請加入我們的細分方向交流群，目前主要有3D視覺、CV&深度學習、SLAM、三維重建、點雲后處理、自動駕駛、CV入門、三維測量、VR/AR、3D人臉識別、醫療影像、缺陷檢測、行人重識別、目標跟蹤、視覺產品落地、視覺競賽、車牌識別、硬件選型、學術交流、求職交流、ORB-SLAM系列源碼交流、深度估計等微信群。一定要備注：研究方向+學校/公司+昵稱，例如：”3D視覺 + 上海交大 + 靜靜“。請按照格式備注，可快速被通過且邀請進群。原創投稿也請聯系。▲長按加微信群或投稿▲長按關注公眾號
3D視覺從入門到精通知識星球：針對3D視覺領域的知識點匯總、入門進階學習路線、最新paper分享、疑問解答四個方面進行深耕，更有各類大廠的算法工程人員進行技術指導。與此同時，星球將聯合知名企業發布3D視覺相關算法開發崗位以及項目對接信息，打造成集技術與就業為一體的鐵桿粉絲聚集區，近2000星球成員為創造更好的AI世界共同進步，知識星球入口：
學習3D視覺核心技術，掃描查看介紹，3天內無條件退款 圈里有高質量教程資料、可答疑解惑、助你高效解決問題覺得有用，麻煩給個贊和在看~