Accurate 3D action recognition using learning on the Grassmann manifold

By Rim Slama, Hazem Wannous, Mohamed Daoudi, Anuj Srivastava. 2015

200 citations recorded by the source corpus.

Subfields: Biomedical Engineering, Computer Vision and Pattern Recognition, 3D Vision & Pose, Human-Computer Interaction

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Biomedical Engineering

1. Jun Liu
2. Wanqing Li
3. Pichao Wang
4. Yonghong Hou
5. Amir Shahroudy
6. G. Wang
7. Chuankun Li
8. Ennio Gambi

In Computer Vision and Pattern Recognition

1. Jun Liu
2. Wanqing Li
3. Pichao Wang
4. G. Wang
5. Yonghong Hou
6. Amir Shahroudy
7. Alex ChiChung Kot
8. Hossein Rahmani

In 3D Vision & Pose

1. Jun Liu
2. Amir Shahroudy
3. Pichao Wang
4. Yonghong Hou
5. Wanqing Li
6. Gang Wang
7. Ennio Gambi
8. Susanna Spinsante

In Human-Computer Interaction

1. Pichao Wang
2. Wanqing Li
3. Jun Liu
4. Yonghong Hou
5. Chuankun Li
6. Amir Shahroudy
7. Liliana Lo Presti
8. Marco La Cascia

Citations in the corpus

• Space-Time Representation of People Based on 3D Skeletal Data: A Review (2016, methodology): “[183] introduced a multi-stage method based on a Grassmann manifold.”
• Skeleton-Based Action Recognition Using Spatio-Temporal LSTM Network with Trust Gates (2017, neutral): “Grassmann Manifold [81] Geometric 88.”
• Spatio-Temporal LSTM with Trust Gates for 3D Human Action Recognition (2016, methodology): “Accuracy Histogram of 3D Joints [5] 90.9% Grassmann Manifold [65] 88.5% Riemannian Manifold [66] 91.5% ST-LSTM (Joint Chain) 91.0% ST-LSTM (Tree) 92.4% ST-LSTM (Tree) + Trust Gate 97.0% Finally, we evaluate the classification performance on early stopping conditions by feeding the first p (0 p 1)…”
• Action Recognition Based on Joint Trajectory Maps Using Convolutional Neural Networks (2016, methodology): “ction of the body joint locations over time. The local 3D body pose is characterized by the current joint locations and differential properties like speed and acceleration of the joints. Slama et al. [68] represented each action sequence as a linear dynamic system that produces 3D joint trajectories”
• Learning features combination for human action recognition from skeleton sequences (2017, methodology): “…4.18% — — [18] — — — 82.0% [21] 88.89% — — — [4] 92.1% 87.28% ± 2.41% 91.5% 87.04% [11] 95.62% — — — [23] 89.48% — 97.08% 90.88% [28] 93.09% — — — [20] 91.21% — 88.5% — [22] 92.03% — — — [8] 92 • In the feature aggregation stage, if the PCA or the power law normalization is turned off, the…”
• Global Context-Aware Attention LSTM Networks for 3D Action Recognition (2017, unknown): “5% Grassmann Manifold [41] 88.”
• A Human Activity Recognition System Using Skeleton Data from RGBD Sensors (2016, background): “Slama et al. [32] exploited the geometric structure of the Grassmann manifold for action analysis.”
• WALKING WALKing walking: Action Recognition from Action Echoes (2017, methodology): “On the other hand, dynamics-based methods treat skeleton data as 3D trajectories of body joints [Lo Presti et al. , 2015; Slama et al. , 2015].”
• Action Recognition Based on Joint Trajectory Maps with Convolutional Neural Networks (2016, background): “Slama et al. [68] represented each action sequence as a linear dynamic system that produces 3D joint trajectories.”
• 3D skeleton-based human action classification: A survey (2016, methodology): “The LTBSVM in [65] describes an action as a collection of time series of 3D locations of the joints.”