Multiple Tree Models for Occlusion and Spatial Constraints in Human Pose Estimation

By Yang Wang, Greg Mori. 2008

171 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Control and Systems Engineering

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

In Control and Systems Engineering

Citations in the corpus

Realtime Multi-person 2D Pose Estimation Using Part Affinity Fields (2016, background): “…models [10, 30, 4, 3, 26, 40, 17], which parametrically encode the spatial relationship between adjacent parts following a kinematic chain, or non-tree models [37, 31, 20, 18, 8] that augment the tree structure with additional edges to capture occlusion, symmetry, and long-range relationships.”
Convolutional Pose Machines (2016, background): “Non-tree models [8, 16, 19, 33, 42] incorporate interactions that introduce loops to augment the tree structure with additional edges that capture symmetry, occlusion and long-range relationships.”
Articulated Human Detection with Flexible Mixtures of Parts (2013, neutral): “Loopy constraints address this limitation but require approximate inference strategies such as sampling [1], [16], [17], loopy belief propagation [18], or iterative approximations [19].”
Pose Machines: Articulated Pose Estimation via Inference Machines (2014, methodology): “Pictorial structure models with non-tree interactions have been employed [12,13,14,15] to estimate pose in a single image.”
Multi-source Deep Learning for Human Pose Estimation (2014, methodology): “On the other hand, many recent works use local body parts [49, 51, 10, 55, 8, 12, 46, 39, 1, 40, 20, 44, 52] in order to handle the many degrees of freedom in body part articulation.”
Better Appearance Models for Pictorial Structures (2009, background): “In many works the model structure E is a tree [10, 11, 21, 22, 23], which enables exact inference, though some works have explored more complex topologies [6, 7, 18, 26, 27, 30].”
Learning hierarchical poselets for human parsing (2011, background): “Examples include common-factor models [12], loopy graphs [9, 18, 25, 26], mixtures of trees [28].”
Multi-task Recurrent Neural Network for Immediacy Prediction (2015, methodology): “Both holistic models [28, 11, 19, 20, 33] and local models [31, 8, 39, 6, 9, 30, 24, 10, 26, 14, 29, 35, 25, 1, 21, 4] were used for estimating the pose of a single person.”
Combining local appearance and holistic view: Dual-Source Deep Neural Networks for human pose estimation (2015, methodology): “Following the pictorial-structure model [19, 1], a variety of part-based methods have been developed for human pose estimation [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15].”
Learning effective human pose estimation from inaccurate annotation (2011, methodology): “…further enhancements over the PSM: (i) we incorporate clustering of partial poses to account for images with poses containing occluded parts; (ii) we greatly improve both accuracy and computational efficiency by modeling part appearance using a state-of-theart mixture of linear classifiers [12].”
Recognizing Human-Object Interactions in Still Images by Modeling the Mutual Context of Objects and Human Poses (2012, unknown): “In order to capture more complex human body articulations, some nontree models have also been proposed [30], [31].”
Articulated part-based model for joint object detection and pose estimation (2011, background): “On the other hand, most pose estimators [12, 20, 10, 5, 16, 15, 11] assume that either the object locations, the object scales, or both are predetermined by either a speciﬁc object detector, or given manually.”
Modeling mutual context of object and human pose in human-object interaction activities (2010, background): “The human poses, on the other hand, are usually highly articulated and many body parts are self-occluded.”
Clustered Pose and Nonlinear Appearance Models for Human Pose Estimation (2010, methodology): “Our method builds on the idea of a 'mixture of trees' model [11] [23].”
Human Pose Estimation Using Body Parts Dependent Joint Regressors (2013, methodology): “For instance, several tree models are combined by a boosting procedure in [37], whereas [28] predicts some parameters of the tree model from the image data.”
Body Parts Dependent Joint Regressors for Human Pose Estimation in Still Images (2014, background): “Non-linear regressors, however, can not resolve some ambiguities that are present in human pose estimation by themselves.”
We Are Family: Joint Pose Estimation of Multiple Persons (2010, methodology): “In articulated HPE some previous works model self-occlusions [7–10].”
OpenPose: Realtime Multi-Person 2D Pose Estimation Using Part Affinity Fields (2018, methodology): “The spatial model for articulated pose is either based on tree-structured graphical models [7], [8], [9], [10], [11], [12], [13], which parametrically encode the spatial relationship between adjacent parts following a kinematic chain, or non-tree models [14], [15], [16], [17], [18] that augment the ”