Recognizing Actions in 3D Using Action-Snippets and Activated Simplices

By Chunyu Wang, John Flynn, Yizhou Wang, Alan L. Yuille. 2016

29 citations recorded by the source corpus.

Pose-based action recognition in 3D is the task of recognizing an action (e.g., walking or running) from a sequence of 3D skeletal poses. This is challenging because of variations due to different ways of performing the same action and inaccuracies in the estimation of the skeletal poses. The training data is usually small and hence complex classifiers risk over-fitting the data. We address this task by action-snippets which are short sequences of consecutive skeletal poses capturing the temporal relationships between poses in an action. We propose a novel representation for action-snippets, called activated simplices. Each activity is represented by a manifold which is approximated by an arrangement of activated simplices. A sequence (of action-snippets) is classified by selecting the closest manifold and outputting the corresponding activity. This is a simple classifier which helps avoid over-fitting the data but which significantly outperforms state-of-the-art methods on standard benchmarks.

Subfields: Artificial Intelligence, Biomedical Engineering, Computer Vision and Pattern Recognition

How researchers see this paper

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

In Artificial Intelligence

1. Junsong Yuan
2. Tae-Kyun Kim
3. Zhiyuan Shi
4. Alan L. Yuille
5. Yizhou Wang
6. Chaoqun Weng
7. Chun-yu Wang
8. Junwu Weng

In Biomedical Engineering

1. Tae-Kyun Kim
2. Zhiyuan Shi
3. Alan L. Yuille
4. Chaoqun Weng
5. Yizhou Wang
6. Chun-yu Wang
7. Junwu Weng
8. Junsong Yuan

In Computer Vision and Pattern Recognition

1. Tae-Kyun Kim
2. Junsong Yuan
3. Luc Van Gool
4. Alan L. Yuille
5. Zhiyuan Shi
6. Cordelia Schmid
7. Xiaogang Wang
8. Dacheng Tao

Citations in the corpus

• Skeleton-Based Human Action Recognition With Global Context-Aware Attention LSTM Networks (2017, neutral): “5% Action-Snippets and Activated Simplices [80] 96.”
• Learning and Refining of Privileged Information-Based RNNs for Action Recognition from Depth Sequences (2017, methodology): “Results on Action3D : We follow a similar evaluation protocol from [45, 46].”
• Spatio-Temporal Naive-Bayes Nearest-Neighbor (ST-NBNN) for Skeleton-Based Action Recognition (2017, methodology): “Similar to [26], we also normalize xp and xv to have l2 norm equals to one.”
• Mining 3D Key-Pose-Motifs for Action Recognition (2016, methodology): “We propose instead to use an alternative quantization method, namely the activated simplices method [24].”
• Global Context-Aware Attention LSTM Networks for 3D Action Recognition (2017, unknown): “5% Action-Snippets and Activated Simplices [50] 96.”