Y. Wu, J. Lin, T. Huang. "Capturing natural hand articulation". Proc. Intl Conf. on Computer Vision (ICCV'01), (2001).  Home/Search   Document Details and Download   Summary   Related Articles   Check

....27] to minimize differences between the silhouette of the projected model and the data. Others [31, 25] have used stereo to avoid occlusions. Appearance based approaches estimate hand postures directly from the images after learning the mapping from image feature space to hand configuration space [30, 29]. Such systems are more useful for recognizing discrete hand states than for general purpose hand tracking. The study of motion and shading together has been formalized [20, 22] recently and extended to multiple views [21] Our approach is model based and hence can work with a single view. Our ....

Ying Wu; Lin, J.Y.; Huang, T.S. Capturing natural hand articulation. ICCV '01 (II:426-432)

....for cluttered images. Our method combines geometric and saliency criteria in a probabilistic way, and identifies line matches that are the least likely to have occurred by chance. 2. Related Work Computer vision systems that estimate 3D hand pose typically do it in the context of tracking [8, 16, 19, 20, 24]. In that context, the pose can be estimated at the current frame as long as the system knows the pose at the previous frame. Since such trackers rely on knowledge about the previous frame, they need to be manually initialized, and cannot recover when they lose the track. Modules that can ....

Y. Wu, J. Lin, and T. Huang. Capturing natural hand articulation. In ICCV, volume 2, pages 426--432, 2001.

....If we control the example generation process, the required property can be achieved by filling the space, storing an example in every node on an R grid in parameter space. This becomes infeasible very quickly as the dimension of 0 increases. Alternatively, it has been observed or conjectured [13, 19] that images of many real world phenomena do not fill the space uniformly, but rather belong to an intrinsically low dimensional subspace, or manifold, and densely covering that subspace is enough to ensure the third property. The last assumption is perhaps the most limiting. It implies that ....

Y. Wu, J. Y. Lin, and T. S. Huang. Capturing natural hand articulation. In Proceedings of the International Conference on Computer Vision, pages 426432, Vancouver, BC, 2001. 11

....and analysis of biological motion have become active research topics in recent years. A common approach to this task is to model the body as a kinematic tree, and reformulate the problem as an articulated body tracking[8] Most of the state of the art systems rely on predefined kinematic models [21, 20, 18]. Some methods require manual initialization, while other use heuristics [15, 7] or predefined protocols [13] to adapt the model to observations. We are interested in a principled way to recover articulated models from observations. The recovered models may then be used for further tracking ....

....structured as follows: in Section 2 we describe relevant prior work, we then describe the probabilistic formulation in Section 3, and finally we present the algorithm used for computations (Section 4) our experiments and the conclusions. 2. Prior Work While state of the art tracking algorithms [21, 9, 5, 20, 18] do not address either model creation or model initialization, the necessity of automating these two steps has been long recognized. The approach in [13] required a subject to follow a set of predefined movements, and recovered the descriptions of body parts and body topology from deformations of ....

Ying Wu, Jonh Y. Lin, and Thomas S. Huang. Capturing natural hand articulation. In Proc. International Conference on Computer Vision, 2001.

....recognition, virtual reality, biometrics, and motion capture. Currently, systems requiring accurate 3D hand parameters tend to use magnetic tracking devices and other non vision based methods [15, 16, 20] Computer vision systems that estimate 3D hand pose do it only in the context of tracking [18, 7, 27, 21] . In that context, the pose can be estimated at the current frame as long as the system knows the pose in the previous frame. The limitation of tracking methods is that they do not address the issue of estimating 3D pose when information about the previous frame is not available. Because of that ....

Ying Wu, J.Y. Lin, and T.S. Huang. Capturing natural hand articulation. In ICCV, volume 2, 21

....recognition, virtual reality, biometrics, and motion capture. Currently, systems requiring accurate 3D hand parameters tend to use magnetic tracking devices and other non vision based methods [7, 8, 11] Computer vision systems that estimate 3D hand pose do it only in the context of tracking [9, 3, 15, 12] . In that context, the pose can be estimated at the current frame as long as the system knows the pose in the previous frame. The limitation of tracking methods is that they do not address the issue of estimating 3D pose when information about the previous frame is not available. Because of that ....

....a hand configuration has 20 degrees of freedom. Grid sampling in such a high dimensional space would be infeasible. Fortunately (at least for our research goals) there is a high degree of correlation in the way those degrees of freedom vary, due to anatomical and behavioral constraints. 10] and [15] applied PCA to hand shapes captured using data gloves, to reduce the dimensionality to eight and seven dimensions respectively. However, even with seven dimensions, discretizing each dimension to, say, five values, would yield about 80,000 configurations, whereas discretizing to 10 values would ....

[Article contains additional citation context not shown here]

Ying Wu, J.Y. Lin, and T.S. Huang. Capturing natural hand articulation. In ICCV, volume 2, pages 426--432, 2001. 14

....the configuration of the joints. However, the algorithm relies on the assumption that all fingertips are visible. Recently, a vision based drawing system was proposed in [16] The 2D shape of a hand is modeled with Bsplines and partitioned sampling is used to track contours in real time. In [26] a model based method for capturing articulated hand motion is presented. The constraints on the joint configurations are learned from natural hand motions, using a data glove as input device. A sequential Monte Carlo tracking algorithm, based on importance sampling, produces good results, but is ....

Y. Wu, L. J. Y., and T. S. Huang. Capturing natural hand articulation. In Proc. 8th Int. Conf. on Computer Vision, volume II, pages 426--432, Vancouver, Canada, July 2001. 6

....sampling is used to track the contours of a 7 DOF model in real time. Partitioned sampling is a method to apply a particle filter to tracking where the configuration space is high dimensional. The idea is to partition the state space, effectively decoupling the motions of different fingers. In [45] Wu et al. present a model based method for capturing articulated hand motion. The constraints on the joint configurations are learned from natural hand motions, using a data glove as input device. It is found that natural hand articulation is highly constrained and that the dimensionality of the ....

Y. Wu, Lin J. Y., and T. S. Huang. Capturing natural hand articulation. In Proc. 8th Int. Conf. on Computer Vision, volume II, pages 426--432, Vancouver, Canada, July 2001.

No context found.

Y. Wu, J. Lin, T. Huang. "Capturing natural hand articulation". Proc. Intl Conf. on Computer Vision (ICCV'01), (2001).

No context found.

Y. Wu, J. Y. Lin, and T. S. Huang. Capturing natural hand articulation. In Proc. 8th Int. Conf. on Computer Vision, volume II, pages 426--432, Vancouver, Canada, July 2001.

No context found.

Wu, Y., Lin, J. & Huang, T. (2001). Capturing natural hand articulation. In IEEE International Conference on Computer Vision (pp. II: 426--432).

No context found.

Y. Wu, J. Y. Lin, and T. S. Huang. Capturing natural hand articulation. pages 426--432, Vancouver, BC, 2001. 8

No context found.

Ying Wu, J.Y. Lin, and T.S. Huang. Capturing natural hand articulation. In ICCV, volume 2, pages 426--432, 2001.

No context found.

Y. Wu, J.Y. Lin, and T.S. Huang. Capturing natural hand articulation. In ICCV, volume 2, pages 426--432, 2001.

No context found.

Y. Wu, L. J. Y., and T. S. Huang. Capturing natural hand articulation. In Proc. 8th Int. Conf. on Computer Vision, volume II, pages 426--432, Vancouver, Canada, July 2001. 6

No context found.

Y. Wu, J. Lin, and T. S. Huang, "Capturing natural hand articulation," in Proc. IEEE Int'l Conference on Computer Vision, vol. II, July 2001, pp. 426--432.

No context found.

Y. Wu, J. Y. Lin, and T. S. Huang. Capturing natural hand articulation. In Proc. 8th International Conference on Computer Vision, pages 426--432, Vancouver, Canada, 2001. IEEE Comp.Soc., IEEE Comp.Soc. Press.

No context found.

Y. Wu, J. Lin, and T. S. Huang. Capturing natural hand articulation. In Proc. IEEE Int'l Conference on Computer Vision, volume II, pages 426--432, Vancouver, July 2001. 7

No context found.

Y. Wu, J. Y. Lin, and T. S. Huang. Capturing natural hand articulation. pages 426--432, Vancouver, BC, 2001. 8

No context found.

Y. Wu, J. Y. Lin, and T. S. Huang. Capturing natural hand articulation. In Proc. 8th International Conference on Computer Vision, pages 426--432, Vancouver, Canada, 2001. IEEE Comp.Soc., IEEE Comp.Soc. Press.

No context found.

Y. Wu, J. Y. Lin, and T. S. Huang. Capturing natural hand articulation. In Proc. 8th Int. Conf. on Computer Vision, volume II, pages 426--432, Vancouver, Canada, July 2001.

No context found.

Y. Wu, J. Y. Lin, and T. S. Huang. Capturing natural hand articulation. In Proc. 8th Int. Conf. on Computer Vision, volume II, 426--432, Vancouver, Canada, July 2001.

No context found.

Y. Wu, J. Y. Lin, and T. S. Huang. Capturing natural hand articulation. pages 426--432, Vancouver, BC, 2001. 8

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC