Tracking loose-limbed people

Sigal, Leonid; Bhatia, Shalini; Roth, Stefan; Black, Michael J.; Isard, Michael

doi:10.1109/cvpr.2004.1315063

Cited by 294 publications

(343 citation statements)

References 24 publications

Supporting

Mentioning

343

Contrasting

Order By: Relevance

“…Articulated motion for the i th joint is estimated in terms of all the (i − 1) joints using twists and exponential maps. Sigal et al in [3], [42] used conditional probability distributions, which can be interpreted as soft articulation constraints, to model the relationships between body joints. Demirdjian et al in [43] imposed kinematic constraints using a linear manifold estimated from the previous body pose.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Linearized Motion Estimation for Articulated Planes

Datta

Sheikh

Kanade

2011

IEEE Trans. Pattern Anal. Mach. Intell.

View full text Add to dashboard Cite

Abstract-In this paper, we describe the explicit application of articulation constraints for estimating the motion of a system of articulated planes. We relate articulations to the relative homography between planes and show that these articulations translate into linearized equality constraints on a linear least squares system, which can be solved efficiently using a Karush-Kuhn-Tucker system. The articulation constraints can be applied for both gradient-based and feature-based motion estimation algorithms and to illustrate we describe a gradient-based motion estimation algorithm for an affine camera and a feature-based motion estimation algorithm for a projective camera that explicitly enforce articulation constraints. We show that explicit application of articulation constraints leads to numerically stable estimates of motion. The simultaneous computation of motion estimates for all the articulated planes in a scene allows us to handle scene areas where there is limited texture information and areas that leave the field of view. Our results demonstrate the wide applicability of the algorithm in a variety of challenging real world cases such as human body tracking, motion estimation of rigid, piecewise planar scenes and motion estimation of triangulated meshes.

show abstract

Section: Related Workmentioning

confidence: 99%

“…Unlike previous constraints [3], [9], we define exact equality constraints on the motion model of the articulated planes for an affine camera (Section 4.1). Articulation constraints can be used for motion estimation in both gradient-based and featurebased motion estimation algorithms.…”

Section: Introductionmentioning

confidence: 99%

Linearized Motion Estimation for Articulated Planes

Datta

Sheikh

Kanade

2011

IEEE Trans. Pattern Anal. Mach. Intell.

View full text Add to dashboard Cite

show abstract

“…However, we prefer to use quaternions because they can be compared using the Euclidean distance metric. Moreover, if the relative rotations between connected parts in the kinematic tree are small, i.e., in the range between 0 and π then we can reduce quaternion representation to 3-tuple instead of 4-tuple, see [21] or [5] for details.…”

Section: Human Motion Trackingmentioning

confidence: 99%

Articulated tracking with manifold regularized particle filter

Gonczarek

Tomczak

2016

Machine Vision and Applications

View full text Add to dashboard Cite

In this paper, we investigate articulated human motion tracking from video sequences using Bayesian approach. We derive a generic particle-based filtering procedure with a low-dimensional manifold. The manifold can be treated as a regularizer that enforces a distribution over poses during tracking process to be concentrated around the lowdimensional embedding. We refer to our method as manifold regularized particle filter. We present a particular implementation of our method based on back-constrained gaussian process latent variable model and gaussian diffusion. The proposed approach is evaluated using the real-life benchmark dataset HumanEva. We show empirically that the presented sampling scheme outperforms sampling-importance resampling and annealed particle filter procedures.

show abstract

“…The temporal prior is incorporated as a state transition matrix, and then the tracking task is formulated as a Bayesian estimation problem. In [15], a single pictorial structure graph model was extended into a dynamic Bayesian network (DBN), where the probabilistic relationships between joints at a given time instant as well as those over time can be learned from motion capture data. Then belief propagation is used as the inference engine to effectively incorporate the top-down spatial prior with bottomup part detection for articulated human tracking.…”

Section: Related Workmentioning

confidence: 99%

“…Given a segmentation Z formed in a location, we represent the boundary of Z by Γ(Z), and the new evaluation score for Z is given by (15) where the first term is the chamfer distance indicating the shape similarity between Γ(Z) and the off-line learned edge prior M, and the second term is the SAD (Sum of absolute differences) that reflects the degree of match between the online learned template P N and I (N +1) . ζ balances the relative importance between the off-line and online learned part priors.…”

Section: Part Detectionmentioning

confidence: 99%