Inferring 3D body pose from silhouettes using activity manifold learning

By Ahmed M. Elgammal, Chan-Su Lee. Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2004. CVPR 2004., 2004

491 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/41c9ec34255b02dab25407b15b4933e53e1e300c

How researchers see this paper

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

Citations in the corpus

Machine Recognition of Human Activities: A Survey (2008, methodology): “Accurate tracking algorithms need to be employed for obtaining motion trajectories [6].”
Robust Estimation of 3D Human Poses from a Single Image (2014, background): “Such holistic representation is able to reduce the ambiguities in the 3D pose estimation and is robust to occlusions ( e.g. missing joints), because it encodes the structural prior of the human skeleton manifold.”
VNect (2017, unknown): “CCS Concepts: • Computing methodologies → Reconstruction ; Computational photography ; Image-based rendering ;”
Vision-based human motion analysis: An overview (2007, background): “Reducing the dimensionality of the pose space is motivated by the observation that human activities are often located on a latent space that is low-dimensional [24,31].”
Single View Human Action Recognition using Key Pose Matching and Viterbi Path Searching (2007, methodology): “Recently, many research efforts have focused on recovering human poses [1] [6] [12], which is considered as a necessary step for view-invariant human action recognition.”
Advances in View-Invariant Human Motion Analysis: A Review (2010, methodology): “In addition, a learning-based framework was introduced by Elgammal and Lee [73] for inferring 3-”
3D People Tracking with Gaussian Process Dynamical Models (2006, background): “Estimates are obtained with deterministic optimization, and look remarkably good despite very noisy, missing or erroneous image data and significant occlusions.”
Sparse Representation for 3D Shape Estimation: A Convex Relaxation Approach (2015, unknown): “There are also many alternative approaches for 3D human pose recovery from single images such as using a known articulated skeleton [25], [26], [27], probabilistic graphical models [28], [29], explicit regression [30], [31], to name a few.”
Spatio-temporal Matching for Human Detection in Video (2014, background): “, silhouettes [18]) to 3D pose with different priors [19, 20].”
Progressive Pose Attention Transfer for Person Image Generation (2019, unknown): “We start with the perspective that the images of all the possible poses and views of a certain person constitute a manifold, which is also suggested in previous works [4, 24].”
Learning Joint Top-Down and Bottom-up Processes for 3D Visual Inference (2006, methodology): “The common experience [9] [22] [2] [7] [19] [28] (besides biological arguments) seems to indicate that for speed and robustness it is also useful to consider a diagnostic (discriminative ) model, designed to condition on the observation, not to generate it.”
Incremental nonlinear dimensionality reduction by manifold learning (2006, background): “The utility of manifold learning has been demonstrated in different applications, such as face pose detection [23], [32], face recognition [53], [51], analysis of facial expressions [20], [11], human motion data interpretation [26], gait analysis [20], [19], visualization of fiber traces [8], and wo”
Reconstructing 3D Human Pose from 2D Image Landmarks (2012, background): “For the single image pose recovery task, some of the earliest work is by Lee and Chen [1] who assumed known limb lengths and recovered pose by pruning a binary interpretation tree that enumerates the entire set of conﬁgurations for an articulated body using physical and structural pruning rules and ”
Priors for people tracking from small training sets (2005, methodology): “Scaling of individual data dimensions was introduced by Grochow et al. [6] to account for different variances of different dimensions of the data.”