Kinematic jump processes for monocular 3D human tracking

By Cristian Sminchisescu, Bill Triggs. 2003 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2003. Proceedings., 2003

360 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/2069e28d8e1619093bfa2f00577f8b0d69f6837c

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

Articulated Body Motion Capture by Stochastic Search (2005, background): “More recently they have extended this work explicitly to take into account the particular ambiguities that arise from human kinematics, “scattering” particles into areas of potential ambiguity and therefore making better use of the particle set (Sminchisescu and Triggs, 2003).”
Tracking multiple humans in crowded environment (2004, methodology): “MCMC-based methods have been recently used for many computer vision problems such as image parsing [9] and articulated body tracking [6].”
Sparseness Meets Deepness: 3D Human Pose Estimation from Monocular Video (2015, background): “In this section, the models that describe the relationships between 3D poses, 2D poses and images are introduced.”
Deep Multitask Architecture for Integrated 2D and 3D Human Sensing (2017, background): “This is in principle still prone to local optima as the monocular 3d human pose estimation cost is non-linear and non-convex even under perfect 3d to 2d model-to-image assignment [42].”
Posebits for Monocular Human Pose Estimation (2014, neutral): “There also exist geometric approaches to lifting 2D pose to 3D by enumerating 3D poses [10, 22] or by exploiting priors from MoCap dataset [17].”
Single-Shot Multi-person 3D Pose Estimation from Monocular RGB (2017, background): “Some works split the problem in two: ﬁrst estimate 2D joints and then lift them to 3D [58, 54, 9, 66, 35, 69, 1, 52, 20, 32, 6, 26, 38, 57, 2], e.g ., by database matching, neural network regression, or ﬁtting the SMPL body model [30].”
3D human pose from silhouettes by relevance vector regression (2004, methodology): “Model-based approaches presuppose an explicitly known parametric body model, and estimate the pose either by directly inverting the kinematics (which requires known image positions for each body part) [15], or by numerically optimizing some form of model-image correspondence metric over the pose var”
MonoCap: Monocular Human Motion Capture using a CNN Coupled with a Geometric Prior (2017, neutral): “F”
Efficient mean-shift tracking via a new similarity measure (2005, methodology): “Knowledge-based tracking algorithms incorporate a priori information about the tracked objects to obtain representations such as projected shape, skin complexion, body blobs, kinematic skeletons and silhouettes [36, 34, 5, 30, 7].”
Structured Prediction of 3D Human Pose with Deep Neural Networks (2016, methodology): “This has been done by explicitly enforcing physical constraints at test time [25, 29] and by data-driven priors over the pose space [6, 28, 33].”
Recovering Accurate 3D Human Pose in the Wild Using IMUs and a Moving Camera (2018, background): “Several works lift 2D detections to 3D using learning or geometric reasoning [18, 29, 35, 9, 26, 49, 33, 48, 44, 34, 13, 43, 45].”
Video Based Reconstruction of 3D People Models (2018, background): “The reason is that pose differences induce additional 3D ambiguities which cannot be uniquely decoupled from global size, even on multiple frames [67, 61, 49].”
Monocular 3D Pose and Shape Estimation of Multiple People in Natural Scenes: The Importance of Multiple Scene Constraints (2018, background): “The problem of minimizing θ and β for monocular error functions, defined over distances between sparse sets of joints, is its ambiguity, as the system is clearly underdetermined, especially for depth related state space directions that couple along camera’s ray of sight [33].”
Iterated Second-Order Label Sensitive Pooling for 3D Human Pose Estimation (2014, background): “Many of the early approaches to 3D human pose estimation employed generative models to search the state space in order to identify configurations of a body model that would best align with image features[12, 32, 29, 34, 15].”
View independent human body pose estimation from a single perspective image (2004, background): “In [11], Sminchisescu and Triggs report an approach for monocular 3D human tracking.”
Fast multiple object tracking via a hierarchical particle filter (2005, neutral): “Model-based tracking algorithms incorporate a priori information about the objects to develop representations such as skin complexion, body blobs, kinematic skeleton, silhouettes or layer information [38, 36, 5, 33, 32, 6].”
MCMC-based particle filtering for tracking a variable number of interacting targets (2005, methodology): “Due to the limitations of importance sampling in highdimensional state spaces, researchers have applied Markov chain Monte Carlo (MCMC) methods to tracking problems [27], [28].”