1.By Object

1.1 Human Body

• Learning to Estimate 3D Human Pose and Shape From a Single Color Image
• Recognizing Human Actions as the Evolution of Pose Estimation Maps
• Human Pose Estimation With Parsing Induced Learner
• Monocular 3D Pose and Shape Estimation of Multiple People in Natural Scenes - The Importance of Multiple Scene Constraints
• Jointly Optimize Data Augmentation and Network Training: Adversarial Data Augmentation in Human Pose Estimation
• V2V-PoseNet: Voxel-to-Voxel Prediction Network for Accurate 3D Hand and Human Pose Estimation From a Single Depth Map
• PoseTrack: A Benchmark for Human Pose Estimation and Tracking
• Cascaded Pyramid Network for Multi-Person Pose Estimation
• Ordinal Depth Supervision for 3D Human Pose Estimation
• Through-Wall Human Pose Estimation Using Radio Signals
• Learning Monocular 3D Human Pose Estimation From Multi-View Images

1.2 Hands

• First-Person Hand Action Benchmark With RGB-D Videos and 3D Hand Pose Annotations
• Depth-Based 3D Hand Pose Estimation: From Current Achievements to Future Goals
• Dense 3D Regression for Hand Pose Estimation
• Gesture Recognition: Focus on the Hands
• Hand PointNet: 3D Hand Pose Estimation Using Point Sets
• Cross-Modal Deep Variational Hand Pose Estimation
• Augmented Skeleton Space Transfer for Depth-Based Hand Pose Estimation
• GANerated Hands for Real-Time 3D Hand Tracking From Monocular RGB

1.3 Others

• Detect-and-Track: Efficient Pose Estimation in Videos
• “檢測和跟蹤：視訊中的高效姿態估計”
• Feature Mapping for Learning Fast and Accurate 3D Pose Inference From Synthetic Images
• “用於學習的特徵對映從合成影象學習快速且準確的3D姿勢推斷”
• DensePose: Dense Human Pose Estimation in the Wild
• ”密集人體：野外人體姿勢估計“
• 3D Human Pose Estimation in the Wild by Adversarial Learning
• “對抗性學習在野外的人體姿態估計”
• 3D Pose Estimation and 3D Model Retrieval for Objects in the Wild
• “野外物體的三維姿態估計與三維模型檢索”
• RotationNet: Joint Object Categorization and Pose Estimation Using Multiviews From Unsupervised Viewpoints
• “RotationNet：使用來自無監督視點的多檢視的聯合物件分類和姿態估計”
• 2D/3D Pose Estimation and Action Recognition Using Multitask Deep Learning
• “使用多工深度學習的2D / 3D姿態估計和動作識別”
• Learning Pose Specific Representations by Predicting Different Views
• “通過預測不同視角來學習姿勢的具體表現”
• Real-Time Seamless Single Shot 6D Object Pose Prediction
• “實時無縫單射6D物件姿態預測”
• Multi-View Consistency as Supervisory Signal for Learning Shape and Pose Prediction
• “多檢視一致性作為學習形狀和姿態預測的監督訊號”

2.By Task

2.1 Pose Estimation:

• DensePose: Dense Human Pose Estimation In The Wild
• Total Capture: A 3D Deformation Model for Tracking Faces, Hands 
• Weakly Supervised Human Body Part Parsing via Pose-Guided Knowledge Transfer
• Synthesizing Images of Humans in Unseen Poses
• Cascaded Pyramid Network for Multi-Person Pose Estimation

2.2 Video Classification/Action Recognition：

• Non-local Neural Networks
• Appearance-and-Relation Networks for Video Classification
• Optical Flow Guided Feature: A Fast and Robust Motion Representation for Video Action Recognition
• Learning to Localize Sound Source in Visual Scenes
• Towards Universal Representation for Unseen Action Recognition
• Non-Linear Temporal Subspace Representations for Activity Recognition
• Fine-grained Activity Recognition in Baseball Videos(workshop) 
• Learning Latent Super-Events to Detect Multiple Activities in Videos

2.3 Video Understanding：

• What Makes a Video a Video: Analyzing Temporal Information in Video Understanding Models and Datasets（spotlight，Feifei組） 
• What have we learned from deep representations for action recognition?
• A Closer Look at Spatiotemporal Convolutions for Action Recognition
• Rethinking Spatiotemporal Feature Learning For Video Understanding
• On the Integration of Optical Flow and Action Recognition（推薦）
• End-to-End Learning of Motion Representation for Video Understanding（Tencent AI Lab）
• Guess Where? Actor-Supervision for Spatiotemporal Action Localization
• A Unifying Contrast Maximization Framework for Event Cameras, with Applications to Motion, Depth, and Optical Flow Estimation
• Video Representation Learning Using Discriminative Pooling
• Can Spatiotemporal 3D CNNs Retrace the History of 2D CNNs and ImageNet?
• Fast End-to-End Trainable Guided Filter
• Density-aware Single Image De-raining using a Multi-stream Dense Network

3.Related Work

3.1 合成 Synthesis

• Multistage Adversarial Losses for Pose-Based Human Image Synthesis
• “基於姿態的人體影象合成的多級對抗損失”
• Synthesizing Images of Humans in Unseen Poses
• “在看不見的姿勢中合成人類的影象”
• Unsupervised Person Image Synthesis in Arbitrary Poses
• “任意姿勢下的無監督人體影象合成”
• End-to-End Recovery of Human Shape and Pose
• “人體形態和姿勢的端對端恢復”
• Deformable GANs for Pose-Based Human Image Generation
• “用於基於姿勢的人類影象生成的可變形GAN”

3.2 相機機位 Camera

• GeoNet: Unsupervised Learning of Dense Depth, Optical Flow and Camera Pose
• “GeoNet：密集深度的無監督學習，光流和相機姿勢”
• Hybrid Camera Pose Estimation
• “混合相機姿勢估計”
• Camera Pose Estimation With Unknown Principal Point
• “帶有未知主要點的相機姿態估計”

3.3 人臉 Face

• Super-FAN: Integrated Facial Landmark Localization and Super-Resolution of Real-World Low Resolution Faces in Arbitrary Poses With GANs
• “Super-FAN：整合的人臉地標定位和任意姿勢下的真實世界低解析度人臉的超解析度”
• Disentangling 3D Pose in a Dendritic CNN for Unconstrained 2D Face Alignment
• “在無限制2D面對準的樹突狀CNN中解構三維姿態”
• Joint Pose and Expression Modeling for Facial Expression Recognition
• “面部表情識別的聯合姿態和表情建模”
• Towards Pose Invariant Face Recognition in the Wild
• “面向野外姿態不變的人臉識別”
• Pose-Robust Face Recognition via Deep Residual Equivariant Mapping
• “基於深度殘差等變對映的姿態魯棒人臉識別”
• UV-GAN: Adversarial Facial UV Map Completion for Pose-Invariant Face Recognition
• “UV-GAN：用於姿態不變臉部識別的對抗面部UV對映完成”
• Pose-Guided Photorealistic Face Rotation
• “姿勢指導真實感臉部旋轉”
• Total Capture: A 3D Deformation Model for Tracking Faces, Hands, and Bodies
• “全部捕獲：用於追蹤面部，手部和身體的3D變形模型”

3.4 其他

• Weakly and Semi Supervised Human Body Part Parsing via Pose-Guided Knowledge Transfer
• “通過姿態引導知識轉移進行弱監督和半監督人體部位解析”
• A Certifiably Globally Optimal Solution to the Non-Minimal Relative Pose Problem
• “非最小相對姿態問題的一個可證明的全域性最優解”
• Fight Ill-Posedness With Ill-Posedness: Single-Shot Variational Depth Super-Resolution From Shading
• “以不適當的姿態應對不適應：來自陰影的單發變分深度超解析度”
• Factoring Shape, Pose, and Layout From the 2D Image of a 3D Scene
• “從3D場景的2D影象中分解形狀，姿態和佈局”
• A Pose-Sensitive Embedding for Person Re-Identification With Expanded Cross Neighborhood Re-Ranking
• “擴充套件交叉鄰居重新排序的人員重新識別的姿態敏感嵌入“
• Improving Occlusion and Hard Negative Handling for Single-Stage Pedestrian Detectors
• ”改善單級行人檢測器的遮擋和硬性負面處理“
• End-to-End Learning of Keypoint Detector and Descriptor for Pose Invariant 3D Matching
• ”針對姿態不變三維匹配的關鍵點檢測器和描述符的端到端學習“
• Non-Blind Deblurring: Handling Kernel Uncertainty With CNNs
• “非盲去模糊：用CNN處理核心不確定性”
• Pose Transferrable Person Re-Identification
• “姿態可移動的人員重新識別”
• LSTM Pose Machines
• “LSTM姿勢機器”
• MX-LSTM: Mixing Tracklets and Vislets to Jointly Forecast Trajectories and Head Poses
• “MX-LSTM：混合Tracklets和Vislets來共同預測軌跡和頭部姿勢”
• PoseFlow: A Deep Motion Representation for Understanding Human Behaviors in Videos
• ”PoseFlow：用於理解視訊中人類行為的深層運動表示“
• PoTion: Pose MoTion Representation for Action Recognition
• ”主題：構成動作識別的MoTion表示“
• Analysis of Hand Segmentation in the Wild
• ”野生動物手部分割分析“

References

Author: 王弗蘭克 https://zhuanlan.zhihu.com/p/34604585

Author: 李光睿 https://zhuanlan.zhihu.com/cvpr2018 

Author: 曠視科技 https://zhuanlan.zhihu.com/p/37582402

Author: 夢寐mayshine http://www.cvmart.net/community/article/detail/286?from=groupmessage&appinstall=0