Beyond trees: common-factor models for 2D human pose recovery

By Xiangyang Lan, Dan Huttenlocher. 2005

189 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

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

Realtime Multi-person 2D Pose Estimation Using Part Affinity Fields (2016, methodology): “The spatial model for articulated pose is either based on tree-structured graphical 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 ”
Convolutional Pose Machines (2016, background): “Non-tree models [7, 14, 17, 29, 38] 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, background): “Another approach to eliminating double counting is the use of stronger pose priors [22].”
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.”
Exploring the Spatial Hierarchy of Mixture Models for Human Pose Estimation (2012, methodology): “To model high-order interaction with tractable computational time, recently [10] proposes to use one (global) latent node to model a mixture of different poses.”
Hierarchical Gaussian process latent variable models (2007, methodology): “Do not distribute. for object recognition (Felzenszwalb & Huttenlocher, 2000; Io.e & Forsyth, 2001) and human pose estimation (Ramanan & Forsyth, 2003; Sigal et al., 2004; Lan & Huttenlocher, 2005).”
Learning hierarchical poselets for human parsing (2011, background): “Examples include common-factor models [12], loopy graphs [9, 18, 25, 26], mixtures of trees [28].”
We Are Family: Joint Pose Estimation of Multiple Persons (2010, methodology): “the upper arms must be attached to the torso) and, in a few works, also other relations such as self-occlusion constraints [7] or the coordination between parts [17] (e.”
Articulated pose estimation with flexible mixtures-of-parts (2011, background): “Another approach to tackling the double-counting phenomena is the use of stronger pose priors, advocated by [17].”
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 ”
Clustered Pose and Nonlinear Appearance Models for Human Pose Estimation (2010, background): “Work related to ours has attempted to address these issues by using a mixture of trees [11] prior or a latent 'pose' variable [24].”
Total Capture: 3D Human Pose Estimation Fusing Video and Inertial Sensors (2017, background): “Lan [18] provide a top down model based approach, considering the conditional independence of parts; however Inter-Limb dependencies (e.”
People-tracking-by-detection and people-detection-by-tracking (2008, background): “But for particular object categories, such as walking people, we can introduce auxiliary state variables that represent the articulation state or an aspect of the object, such as different phases in the walking cycle of a person [12], and make the parts conditionally independent.”
Learning Joint Top-Down and Bottom-up Processes for 3D Visual Inference (2006, background): “Systems based on global [17, 4, 31, 17] or part-based human detectors [19, 21, 15] have been demonstrated convincingly, and can be an alternative 2d front end to the integrated 3d reconstruction scheme proposed here.”
Guiding model search using segmentation (2005, background): “However, previous work [9, 21] has shown that tree models fail to capture some additional dependencies between body parts.”
Pictorial structures revisited: People detection and articulated pose estimation (2009, background): “It is certainly possible to incorporate action specific constraints into the prior, and to combine them with articulation dynamics to enable tracking, as is done for example in [10] [23].”
Measure Locally, Reason Globally: Occlusion-sensitive Articulated Pose Estimation (2006, background): “Recent approaches to articulated human body detection and pose estimation exploit part-based tree-structured models [3, 5, 8, 13, 15, 17] that capture kinematic relations between body parts.”
HumanEva: Synchronized Video and Motion Capture Dataset for Evaluation of Articulated Human Motion (2006, methodology): “As a convention from previous methods [3, 15, 37, 36] that have already used this error metric we compute the 3D error in millimeters ( mm ) and 2D error directly in the image in pixels ( pix ).”
A Non-parametric Bayesian Network Prior of Human Pose (2013, unknown): “Despite this flexibility a kinematic tree model has clear limitations, as sharply argued in [15]; it is unable to express the coordination of different limbs and fails to represent global balance and gravity constraints.”
Vision-based human motion analysis: An overview (2007, background): “Trees are extended with correlations between body parts in [50].”