Co-occurrence Feature Learning for Skeleton based Action Recognition using Regularized Deep LSTM Networks
arXiv:1603.07772
Abstract
Skeleton based action recognition distinguishes human actions using the trajectories of skeleton joints, which provide a very good representation for describing actions. Considering that recurrent neural networks (RNNs) with Long Short-Term Memory (LSTM) can learn feature representations and model long-term temporal dependencies automatically, we propose an end-to-end fully connected deep LSTM network for skeleton based action recognition. Inspired by the observation that the co-occurrences of the joints intrinsically characterize human actions, we take the skeleton as the input at each time slot and introduce a novel regularization scheme to learn the co-occurrence features of skeleton joints. To train the deep LSTM network effectively, we propose a new dropout algorithm which simultaneously operates on the gates, cells, and output responses of the LSTM neurons. Experimental results on three human action recognition datasets consistently demonstrate the effectiveness of the proposed model.
AAAI 2016 conference
References in corpus (2)
Cited by in corpus (37)
- An End-to-End Spatio-Temporal Attention Model for Human Action Recognition from Skeleton Data
- NTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysis
- PKU-MMD: A Large Scale Benchmark for Continuous Multi-Modal Human Action Understanding
- MS-ASL: A Large-Scale Data Set and Benchmark for Understanding American Sign Language
- A Review of Computational Approaches for Evaluation of Rehabilitation Exercises
- Semantics-Guided Neural Networks for Efficient Skeleton-Based Human Action Recognition
- Human Action Performance using Deep Neuro-Fuzzy Recurrent Attention Model
- Learning stochastic differential equations using RNN with log signature features
- Molecular Property Prediction: A Multilevel Quantum Interactions Modeling Perspective
- DanceIt: Music-inspired Dancing Video Synthesis
- Simultaneous Joint and Object Trajectory Templates for Human Activity Recognition from 3-D Data
- Optimized Skeleton-based Action Recognition via Sparsified Graph Regression
- Making the Invisible Visible: Action Recognition Through Walls and Occlusions
- Self-Attention Network for Skeleton-based Human Action Recognition
- Skeleton Focused Human Activity Recognition in RGB Video
- Shifted Chunk Transformer for Spatio-Temporal Representational Learning
- NeurReg: Neural Registration and Its Application to Image Segmentation
- Human Action Recognition with Multi-Laplacian Graph Convolutional Networks
- A-MAL: Automatic Movement Assessment Learning from Properly Performed Movements in 3D Skeleton Videos
- Predictively Encoded Graph Convolutional Network for Noise-Robust Skeleton-based Action Recognition
- Skeleton Based Action Recognition using a Stacked Denoising Autoencoder with Constraints of Privileged Information
- Skeleton-Based Online Action Prediction Using Scale Selection Network
- iMiGUE: An Identity-free Video Dataset for Micro-Gesture Understanding and Emotion Analysis
- SpeechNAS: Towards Better Trade-off between Latency and Accuracy for Large-Scale Speaker Verification
- An Emerging Coding Paradigm VCM: A Scalable Coding Approach Beyond Feature and Signal
- Coarse Temporal Attention Network (CTA-Net) for Driver's Activity Recognition
- Learning Coupled Spatial-temporal Attention for Skeleton-based Action Recognition
- Contrastive Video Representation Learning via Adversarial Perturbations
- Totally Deep Support Vector Machines
- Action Recognition with Kernel-based Graph Convolutional Networks
- Literature Review of Action Recognition in the Wild
- Automatic Remaining Useful Life Estimation Framework with Embedded Convolutional LSTM as the Backbone
- TrajectoryNet: a new spatio-temporal feature learning network for human motion prediction
- DeepActsNet: Spatial and Motion features from Face, Hands, and Body Combined with Convolutional and Graph Networks for Improved Action Recognition
- GeomNet: A Neural Network Based on Riemannian Geometries of SPD Matrix Space and Cholesky Space for 3D Skeleton-Based Interaction Recognition
- In Defense of LSTMs for Addressing Multiple Instance Learning Problems
- DWnet: Deep-Wide Network for 3D Action Recognition