Awesome

Papers

RNN

NTU

[1] NTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysis

• intro: CVPR 2016, NTU RGB-D 60 Dataset, Part-Aware LSTM, Benchmark Evaluation
• arxiv: https://arxiv.org/abs/1604.02808
• github: https://github.com/shahroudy/NTURGB-D
• github(TF): https://github.com/FesianXu/PLSTM

[2] Spatio-Temporal LSTM with Trust Gates for 3D Human Action Recognition

• intro: T-PAMI 2017, Spatio-Temporal LSTM
• arxiv: https://arxiv.org/abs/1607.07043
• github: https://github.com/kinect59/Spatio-Temporal-LSTM

[3] Skeleton Based Human Action Recognition with Global Context-Aware Attention LSTM Networks

• intro: CVPR 2017, Attention mechanism
• arxiv: https://arxiv.org/abs/1707.05740

[4] NTU RGB+D 120: A Large-Scale Benchmark for 3D Human Activity Understanding

• intro: TPAMI 2019, NTU RGB-D 120 Dataset, One shot 3D Action Recognition
• dataset:[http://rose1.ntu.edu.sg/datasets/actionrecognition.asp](NTU RGB-D 120)
• arxiv: https://arxiv.org/pdf/1905.04757.pdf
• github: https://github.com/shahroudy/NTURGB-D

[5] Modeling Temporal Dynamics and Spatial Configurations of Actions Using Two-Stream Recurrent Neural Networks

• intro: CVPR 2017, Temporal RNN(Stacked RNN and Hierarchical RNN), Spatial RNN(Chain sequence and Traversal sequence)
• arxiv: https://arxiv.org/abs/1704.02581
• github: https://github.com/hongsong-wang/RNN-for-skeletons

[6] Learning content and style: Joint action recognition and person identification from human skeletons

• intro: PR 2018, Multi-task learning about action recognition and person identification
• github: https://github.com/hongsong-wang/Beyond-Joints

[7] Ensemble Deep Learning for Skeleton-based Action Recognition using Temporal Sliding LSTM networks

• intro: ICCV 2017, Refine NTU-D dataset
• github: https://github.com/InwoongLee/TS-LSTM

[8] View Adaptive Recurrent Neural Networks for High Performance Human Action Recognition from Skeleton Data

• intro: ICCV 2017, View Adaptation Subnetwork (RNN)
• arxiv: https://arxiv.org/abs/1703.08274

[9] View Adaptive Neural Networks for High Performance Skeleton-based Human Action Recognition

• intro: View Adaptation Subnetwork (RNN and CNN)
• arxiv: http://cn.arxiv.org/abs/1804.07453

[10] Independently Recurrent Neural Network (IndRNN): Building A Longer and Deeper RNN

• intro: CVPR 2018, Introduce the residual connectins to RNN
• arxiv: http://cn.arxiv.org/abs/1803.04831
• github(TF): https://github.com/batzner/indrnn
• github(Pytorch): https://github.com/StefOe/indrnn-pytorch

[11] Skeleton-Based Action Recognition with Spatial Reasoning and Temporal Stack Learning

• intro: ECCV 2018, Spatial reasoning network, Temporal stack learning network
• arxiv: https://arxiv.org/abs/1805.02335v1

[12] Adding Attentiveness to the Neurons in Recurrent Neural Networks

• intro: ECCV 2018, Element-wise-Attention Gate for an RNN Block
• arxiv: https://arxiv.org/abs/1807.04445v1

CNN

Hikvision

[1] Skeleton-based Action Recognition with Convolutional Neural Networks

• intro: ICMEW 2017, Two stream cnn, Transformer
• arxiv: https://arxiv.org/abs/1704.07595

[2] Co-occurrence Feature Learning from Skeleton Data for Action Recognition and Detection with Hierarchical Aggregation

• intro: IJCAI 2018, Hierarchical co-occurrence feature, Transposing
• arxiv: https://arxiv.org/abs/1804.06055

[3] Skeleton-Based Action Recognition with Synchronous Local and Non-local Spatio-temporal Learning and Frequency Attention

• intro: Frequency domain analysis, Non-local operation, Soft-margin focal loss, Transform Network
• arxiv: https://arxiv.org/abs/1811.04237

[4] Three-Stream Convolutional Neural Network with Multi-task and Ensemble Learning for 3D Action Recognition

• intro: CVPRW 2019, Multi-feature CNN, Muti-task and Ensemble

GCN

[1] Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognition

• intro: AAAI 2018, Graph convolutional networks
• arxiv: https://arxiv.org/pdf/1801.07455.pdf
• github: https://github.com/yysijie/st-gcn

[2] Deep Progressive Reinforcement Learning for Skeleton-based Action Recognition

• intro: CVPR 2018, Using reinforcement learning to select frames

[3] Non-Local Graph Convolutional Networks for Skeleton-Based Action Recognition

• intro: Two stream gcn, Non-Local network
• arxiv: https://arxiv.org/abs/1805.07694v2

[4] Actional-Structural Graph Convolutional Networks for Skeleton-based Action Recognition

• intro: CVPR 2019, Actional-Structural GCN
• arxiv: https://arxiv.org/pdf/1904.12659.pdf
• github: https://github.com/limaosen0/AS-GCN

[5] An Attention Enhanced Graph Convolutional LSTM Network for Skeleton-Based Action Recognition

• intro: CVPR 2019, Graph Convolutional LSTM, Attention, Two part
• arxiv: https://arxiv.org/pdf/1902.09130.pdf

[6] Two-Stream Adaptive Graph Convolutional Networks for Skeleton-Based Action Recognition

• github: https://github.com/lshiwjx/2s-AGCN

[7] JOINTS RELATION INFERENCE NETWORK FOR SKELETON-BASED ACTION RECOGNITION

• intro: ICIP 2019, GCN+CNN, Optimal adjacency matrices

Other GITHUB Repos for Skeleton-based Action Recognition Papers and Small Notes

• https://github.com/cagbal/Skeleton-Based-Action-Recognition-Papers-and-Notes
• https://github.com/niais/Awesome-Skeleton-based-Action-Recognition