Feature Tracking and Motion Compensation for Action Recognition

Uemura, Hirofumi; Ishikawa, Seiji; Mikolajczyk, Krystian

doi:10.5244/c.22.30

Cited by 63 publications

(62 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Solution to the problem of detection and monitoring ground moving objects is given in a large number of papers, for example, (Hayman & Eklundh, 2001;Ren et al, 2003;Uemura et al, 2008;Borshukov et al . ;Ke & Kanade, 2001;Tao et al, 2007).…”

Section: Literature Reviewmentioning

confidence: 99%

Ensuring the Accuracy of Traffic Monitoring Using Unmanned Aerial Vehicles Vision Systems

Kim¹,

Bodunkov²,

Cherkezov³

2015

MAS

View full text Add to dashboard Cite

The paper is dedicated to the organization of traffic monitoring using unmanned aerial vehicles (UAV). It is demonstrated that the monitoring of traffic makes high demands on the accuracy of determining the position of the vehicle on the road.It is assumed that the purpose of monitoring is detecting specific situations, which may include accident, in particular, car collision; traffic accidents, reducing the bandwidth of the road section; movement of the vehicle, being a threat to other road users.Detection of such situations requires assessment of the following at the received images:• Vehicles position relatively to the road markings;• Vehicles position relatively to each other;• Vehicles speed. Review of the literature showed that the existing tools for tracking ground objects movements provide sufficiently accurate assessment of the vehicles coordinates at the images. Thus, an important issue is the estimation of vehicle position with respect to the road, i.e. in the ground coordinate system of the road. Different options of the vehicle position assessment relatively the road are researched.Evaluation of the content and accuracy of the standard UAV navigation system showed that the option of monitoring based on the use of UAV position assessment relative to the ground coordinate system and the vehicles is non-implementable because of lack of precision at the standard navigation system, including, corrected using the satellite navigation system. Assessing the position of the vehicle relative to the roadside is proposed to be made using image processing algorithms, particularly the contour lines highlighting and the Hough algorithm for straight segments highlighting.The research shows that this option based on direct assessment of the situation with respect to the vehicle position on the road image is physically implementable.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Ensuring the Accuracy of Traffic Monitoring Using Unmanned Aerial Vehicles Vision Systems

Kim¹,

Bodunkov²,

Cherkezov³

2015

MAS

View full text Add to dashboard Cite

show abstract

“…Uemura et al [23] proposed human action recognition based on the KLT tracker and SIFT descriptor. Multiple interest point detectors were used to provide a large number of interest points for every frame.…”

Section: Trajectories and Variantsmentioning

confidence: 99%

Ordered Trajectories for Large Scale Human Action Recognition

Murthy

Goecke

2013

2013 IEEE International Conference on Computer Vision Workshops

View full text Add to dashboard Cite

Recently, a video representation based on dense trajectories has been shown to outperform other human action recognition methods on several benchmark datasets. In dense trajectories, points are sampled at uniform intervals in space and time and then tracked using a dense optical flow field. The uniform sampling does not discriminate objects of interest from the background or other objects. Consequently, a lot of information is accumulated, which actually may not be useful. Sometimes, this unwanted information may bias the learning process if its content is much larger than the information of the principal object(s) of interest. This can especially escalate when more and more data is accumulated due to an increase in the number of action classes or the computation of dense trajectories at different scales in space and time, as in the Spatio-Temporal Pyramidal approach. In contrast, we propose a technique that selects only a few dense trajectories and then generates a new set of trajectories termed 'ordered trajectories'. We evaluate our technique on the complex benchmark HMDB51, UCF50 and UCF101 datasets containing 50 or more action classes and observe improved performance in terms of recognition rates and removal of background clutter at a lower computational cost.

show abstract

“…In Yuan et al [171] clustered key-point trajectories based on spatial proximities and motion patterns. Like [141], this method extracts relative features from clusters of trajectories on the background that describe the motion differently from those emanating from the foreground. As a result, the effect of camera motion can be alleviated using this approach.…”

Section: Trajectory Descriptorsmentioning

confidence: 99%

“…The velocity histories of key-point trajectories are modeled by Messing et al [87], who observed that velocity information is useful for detecting daily living actions in high-resolution videos. Uemura et al [141] combined feature tracking and frame segmentation to estimate dominant planes in the scene, which were used for motion compensation. In Yuan et al [171] clustered key-point trajectories based on spatial proximities and motion patterns.…”

Section: Trajectory Descriptorsmentioning

confidence: 99%

High-level event recognition in unconstrained videos

Jiang

Bhattacharya

Chang

et al. 2012

Int J Multimed Info Retr

158

106

View full text Add to dashboard Cite

The goal of high-level event recognition is to automatically detect complex high-level events in a given video sequence. This is a difficult task especially when videos are captured under unconstrained conditions by nonprofessionals. Such videos depicting complex events have limited quality control, and therefore, may include severe camera motion, poor lighting, heavy background clutter, and occlusion. However, due to the fast growing popularity of such videos, especially on the Web, solutions to this problem are in high demands and have attracted great interest from researchers. In this paper, we review current technologies for complex event recognition in unconstrained videos. While the existing solutions vary, we identify common key modules and provide detailed descriptions along with some insights for each of them, including extraction and representation of low-level features across different modalities, classification strategies, fusion techniques, etc. Publicly available benchmark datasets, performance metrics, and related research forums are also described. Finally, we discuss promising directions for future research.

show abstract

Feature Tracking and Motion Compensation for Action Recognition

Cited by 63 publications

References 32 publications

Ensuring the Accuracy of Traffic Monitoring Using Unmanned Aerial Vehicles Vision Systems

Ensuring the Accuracy of Traffic Monitoring Using Unmanned Aerial Vehicles Vision Systems

Ordered Trajectories for Large Scale Human Action Recognition

High-level event recognition in unconstrained videos

Contact Info

Product

Resources

About