Estimating 3D Pose via Stochastic Search and Expectation Maximization

Daubney, Ben; Xie, Xianghua

doi:10.1007/978-3-642-14061-7_7

Cited by 4 publications

(6 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A limitation with these approaches is that they are iterative and need convergence for a solution to be found. As noted in [30,31], this convergence happens in a particular order for objects modeled as a kinematic chain. Typically, those parts nearer a fixed node must converge before parts further down the model can do so.…”

Section: Introductionmentioning

confidence: 93%

“…However, as the connection between parts is soft, the model is less constrained and slippage can occur, where two limbs can be joined at a very unlikely location or may not even be physically joined. It has been shown that given a known root location, models that have fixed joint positions outperform loose-limbed models at estimating 3D pose [31].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Fixing the root node: Efficient tracking and detection of 3D human pose through local solutions

Daubney

Xie

Deng

et al. 2016

Image and Vision Computing

Self Cite

View full text Add to dashboard Cite

A B S T R A C T 3D human pose estimation is a very difficult task. In this paper we propose that this problem can be more easily solved by first finding the solutions to a set of easier sub-problems. These are to locally estimate pose conditioned on a fixed root node state, which defines the global position and orientation of the person. The global solution can then be found using information extracted during this procedure. This approach has two key benefits: The first is that each local solution can be found by modeling the articulated object as a kinematic chain, which has far less degrees of freedom than alternative models. The second is that by using this approach we can represent, or support, a much larger area of the posterior than is currently possible. This allows far more robust algorithms to be implemented since there is far less pressure to prune the search space to free up computational resources. We apply this approach to two problems: The first is single frame monocular 3D pose estimation, where we propose a method to directly extract 3D pose without first extracting any intermediate 2D representation or being dependent on strong spatial prior models. The second is multi-view 3D tracking where we show that using the above technique results in an approach that is far more robust than current approaches, without relying on strong temporal prior models. In both domains we demonstrate the strength and versatility of the proposed method.

show abstract

Section: Introductionmentioning

confidence: 93%

Section: Introductionmentioning

confidence: 99%

Fixing the root node: Efficient tracking and detection of 3D human pose through local solutions

Daubney

Xie

Deng

et al. 2016

Image and Vision Computing

Self Cite

View full text Add to dashboard Cite

show abstract

“…Human motion capture and activity recognition have proved viable in, for example, computer graphics, media production, robotics, and video surveillance applications throughout the years [25,20,1,27,5,26,18,7,8], though it still remains an open and challenging problem. There is however already a body of work interested in the detection and recognition of social interaction between multiple people [11,14], which is particularly difficult since the actions of multiple subjects must be inferred and understood.…”

Section: Introductionmentioning

confidence: 99%

A bag of words approach to subject specific 3D human pose interaction classification with random decision forests

2014

Self Cite

View full text Add to dashboard Cite

In this work, we investigate whether it is possible to distinguish conversational interactions from observing human motion alone, in particular subject specific gestures in 3D. We adopt Kinect sensors to obtain 3D displacement and velocity measurements, followed by wavelet decomposition to extract low level temporal features. These features are then generalized to form a visual vocabulary that can be further generalized to a set of topics from temporal distributions of visual vocabulary. A subject specific supervised learning approach based on Random Forests is used to classify the testing sequences to seven different conversational scenarios. These conversational scenarios concerned in this work have rather subtle differences among them. Unlike typical action or event recognition, each interaction in our case contain many instances of primitive motions and actions, many of which are shared among different conversation scenarios. That is the interactions we are concerned with are not micro or instant events, such as hugging and high-five, but rather interactions over a period of time that consists rather similar individual motions, micro actions and interactions. We believe this is among one of the first work that is devoted to subject specific conversational interaction classification using 3D pose features and to show this task is indeed possible.

show abstract

“…According to [17], the main strategies for model-based poseestimation usually are global optimization such as interactive simulated annealing [18], filtering/smoothing/prediction such as Kalman filtering or particle filtering, local optimization such as iterative closest point (ICP) or a combination of these strategies such as in Juergen et al [17]. In [19], they represented the human body as a probability graphical model just as the one in [4,5,6,7,8] but adding some constraints between connected parts and called it Fixed Joint Model (FJM). This new representation helps the convergence of the global sampling process to go faster.…”

Section: Introductionmentioning

confidence: 99%

Video-based analysis of gait with side views

Nguyen

Meunier

2014

2014 4th International Conference on Image Processing Theory, Tools and Applications (IPTA)

View full text Add to dashboard Cite

Gait analysis has emerged as one of the most important medical field recently due to its wide range applications. Many recent computer-vision-based approaches have been made to reduce the cost of the expensive motion capture systems while assuring high accuracy result. In this paper, we introduce our new low-cost gait analysis system that is composed of two cheap monocular cameras (camcorders) placed on the left and right side of a treadmill. Each half human skeleton model is reconstructed from each view based on dynamic color segmentation, the gait analysis is then performed on these two models. We validate our method with ground truth data obtained with markers manually placed on the subject body. Comparison with the 3D skeleton reconstructed with the Kinect sensor are also presented. Our results on different subjects show that although our methodology is simple and low-cost, it is very efficient and provides accurate measure for gait assessment.

show abstract

Estimating 3D Pose via Stochastic Search and Expectation Maximization

Cited by 4 publications

References 12 publications

Fixing the root node: Efficient tracking and detection of 3D human pose through local solutions

Fixing the root node: Efficient tracking and detection of 3D human pose through local solutions

A bag of words approach to subject specific 3D human pose interaction classification with random decision forests

Video-based analysis of gait with side views

Contact Info

Product

Resources

About