Motion Context: A New Representation for Human Action Recognition

Zhang, Ziming; Hu, Yiqun; Chan, Syin; Chia, Liang-Tien

doi:10.1007/978-3-540-88693-8_60

Cited by 110 publications

(85 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Nevertheless, this approach remains limited to offline applications. Other approaches like (Tabia et al, 2012) and (Zhang et al, 2008) use polar space representations to characterize activities. However they compute their descriptors on entire sequences, thus do not explicitly provide on-line recognition capabilities.…”

Section: Evaluation and Resultsmentioning

confidence: 99%

Action-centric Polar Representation of Motion Trajectories for Online Action Recognition

Martínez

Manzanera

Gouiffès

et al. 2016

Proceedings of the 11th Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications

View full text Add to dashboard Cite

Abstract:This work introduces a novel action descriptor that represents activities instantaneously in each frame of a video sequence for action recognition. The proposed approach first characterizes the video by computing kinematic primitives along trajectories obtained by semi-dense point tracking in the video. Then, a frame level characterization is achieved by computing a spatial action-centric polar representation from the computed trajectories. This representation aims at quantifying the image space and grouping the trajectories within radial and angular regions. Motion histograms are then temporally aggregated in each region to form a kinematic signature from the current trajectories. Histograms with several time depths can be computed to obtain different motion characterization versions. These motion histograms are updated at each time, to reflect the kinematic trend of trajectories in each region. The action descriptor is then defined as the collection of motion histograms from all the regions in a specific frame. Classic support vector machine (SVM) models are used to carry out the classification according to each time depth. The proposed approach is easy to implement, very fast and the representation is consistent to code a broad variety of actions thanks to a multi-level representation of motion primitives. The proposed approach was evaluated on different public action datasets showing competitive results (94% and 88.7% of accuracy are achieved in KTH and UT datasets, respectively), and an efficient computation time.

show abstract

Section: Evaluation and Resultsmentioning

confidence: 99%

Action-centric Polar Representation of Motion Trajectories for Online Action Recognition

Martínez

Manzanera

Gouiffès

et al. 2016

Proceedings of the 11th Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications

View full text Add to dashboard Cite

show abstract

“…The method based on spatiotemporal interest point can provide abundant description and expression. In this method [30], spatiotemporal words are composed of extracted spatiotemporal interest points for each segment of video sequence. In complicated videos, they can be expressed by response function of linear filter.…”

Section: Intelligent Behavioral Analysismentioning

confidence: 99%

A Survey on Video Transmission Techniques

Yuan¹,

Yang²

2014

TOEEJ

View full text Add to dashboard Cite

Common video transmission system includes the following work like video collection and encoding of sending end, network transmission and reception, decoding and analysis of receiving end. This paper mainly focuses on elaborating and concluding principles and related techniques involved in video transmission system, such as rate control, transmission control mechanism, video retrieval, intelligent video analysis and error processing of video etc. In conclusion, it points out existing problems of video transmission techniques at present and future development direction of the researches.

show abstract

“…Recently, a common way for representing action is using quantized local features around interest points. In [9], action representation is derived by using 2D quantized local features of motion images word and quantized angles and distances between visual words and the reference point. Liu and Shah [10] investigated the optimal number of video-words by using Maximization of Mutual Information clustering.…”

Section: Related Workmentioning

confidence: 99%

Learning Group Activity in Soccer Videos from Local Motion

Kong

Zhang

et al. 2010

Computer Vision – ACCV 2009

View full text Add to dashboard Cite

Abstract. This paper proposes a local motion-based approach for recognizing group activities in soccer videos. Given the SIFT keypoint matches on two successive frames, we propose a simple but effective method to group these keypoints into the background point set and the foreground point set. The former one is used to estimate camera motion and the latter one is applied to represent group actions. After camera motion compensation, we apply a local motion descriptor to characterize relative motion between corresponding keypoints on two consecutive frames. The novel descriptor is effective in representing group activities since it focuses on local motion of individuals and excludes noise such as background motion caused by inaccurate compensation. Experimental results show that our approach achieves high recognition rates in soccer videos and is robust to inaccurate compensation results.

show abstract

Motion Context: A New Representation for Human Action Recognition

Cited by 110 publications

References 17 publications

Action-centric Polar Representation of Motion Trajectories for Online Action Recognition

Action-centric Polar Representation of Motion Trajectories for Online Action Recognition

A Survey on Video Transmission Techniques

Learning Group Activity in Soccer Videos from Local Motion

Contact Info

Product

Resources

About