Determination of 3D human body postures from a single view

Lee, Hsi-Jian; Chen, Zen

doi:10.1016/0734-189x(85)90094-5

Cited by 154 publications

(19 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This gets further complicated in the articulated case (Lee and Chen 1985;Morris and Rehg 1998;Deutscher et al 2000;Sidenbladh et al 2000;Choo and Fleet 2001;Sminchisescu andTriggs 2002a, 2003;Sminchisescu and Jepson 2004b). One effective way to deal with ambiguity (multivaluedness) arising in perceptual, 3d from 2d inference problems, is to use mixture of experts or their conditional counterparts (Rosales and Sclaroff 2002;.…”

Section: Twin Gaussian Processes (Tgp)mentioning

confidence: 99%

“…Reconstructing articulated 3d pose from monocular model-image point correspondences is certainly ambiguous, according to geometric (Lee and Chen 1985;Morris and Rehg 1998;Sminchisescu and Triggs 2003) and computational studies (Deutscher et al 2000;Sidenbladh et al 2000;Choo and Fleet 2001;Sminchisescu andTriggs 2002a, 2002b;Rosales and Sclaroff 2002;) of the problem. For other image features and volumetric models, the problem may be more difficult to analyze geometrically, but poses that correspond to reflective placements of limbs w.r.t.…”

Section: Twin Gaussian Processes (Tgp)mentioning

confidence: 99%

See 1 more Smart Citation

Twin Gaussian Processes for Structured Prediction

2009

View full text Add to dashboard Cite

We describe twin Gaussian processes (TGP), a generic structured prediction method that uses Gaussian process (GP) priors on both covariates and responses, both multivariate, and estimates outputs by minimizing the Kullback-Leibler divergence between two GP modeled as normal distributions over finite index sets of training and testing examples, emphasizing the goal that similar inputs should produce similar percepts and this should hold, on average, between their marginal distributions. TGP captures not only the interdependencies between covariates, as in a typical GP, but also those between responses, so correlations among both inputs and outputs are accounted for. TGP is exemplified, with promising results, for the reconstruction of 3d human poses from monocular and multicamera video sequences in the recently introduced HumanEva benchmark, where we achieve 5 cm error on average per 3d marker for models trained jointly, using data from multiple people and multiple activities. The method is fast and automatic: it requires no hand-crafting of the initial pose, camera calibration parameters, or the availability of a 3d body model associated with human subjects used for training or testing.

show abstract

Section: Twin Gaussian Processes (Tgp)mentioning

confidence: 99%

Section: Twin Gaussian Processes (Tgp)mentioning

confidence: 99%

Twin Gaussian Processes for Structured Prediction

2009

View full text Add to dashboard Cite

show abstract

“…As far as human body models are concerned, they vary widely in the levels of details. In earlier research, simple stick ÿgure models were frequently used [6,29], in which body parts are represented by sticks connected by joints. More complex volumetric human models, such as cylinder [7,8,30], truncated cone [9,10] and super-quadrics [11], were used in later work.…”

Section: Related Workmentioning

confidence: 99%

People tracking based on motion model and motion constraints with automatic initialization

Ning

Tan

Wang

et al. 2004

Pattern Recognition

View full text Add to dashboard Cite

“…We make a structure model of a human by a stick figure model [7] as illustrated in figure 1. The human body is described by several sticks which represent limbs and are articulated at joints.…”

Section: Structure Modelmentioning

confidence: 99%

Posture estimation using structure and motion models

Iwai

Ogaki²,

Yachida³

1999

Proceedings of the Seventh IEEE International Conference on Computer Vision

View full text Add to dashboard Cite

Sensing of human motion is very important for humancomputer interactive applications such as virtual reality, gesture recognition, and communication. A vision system is suitable for human-computer interaction since this involves passive sensing and the system can estimate the motion of the user without any discomfort for the user. In this paper, we propose an algorithm for fast posture estimation of the user from an image sequence by using the position information of the hands and head. Our algorithm is founded on a model based method. The parameters of a geometric model are calculated from human kinematics. It is inadequate to obtain a unique solution from the image information such as position of the head and hands. The unknown parameters are predicted by the motion models and the previous posture parameters, then the unknown parameters are adjusted by a minimization method.

show abstract

Determination of 3D human body postures from a single view

Cited by 154 publications

References 11 publications

Twin Gaussian Processes for Structured Prediction

Twin Gaussian Processes for Structured Prediction

People tracking based on motion model and motion constraints with automatic initialization

Posture estimation using structure and motion models

Contact Info

Product

Resources

About