Reconstruction of articulated objects from point correspondences in a single uncalibrated image

By Camillo Jose Taylor. 2000

436 citations recorded by the source corpus.

Subfields: Aerospace Engineering, 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 Aerospace Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

End-to-End Recovery of Human Shape and Pose (2017, methodology): “In order to constrain the inherent ambiguity in 2D-to-3D estimation, these methods use various priors [42].”
Exploiting Temporal Information for 3D Human Pose Estimation (2017, methodology): “Jiang [33] used the 2D joint locations to estimate a set of hypothesis 3D poses using Taylor’s algorithm [34] and used them to query a large database of motion capture data to find the nearest neighbor.”
Robust Estimation of 3D Human Poses from a Single Image (2014, methodology): “Elgammal et al.[3] propose to learn view-based silhouettes manifolds and the mapping function from the manifold to 3D poses.”
A survey on still image based human action recognition (2014, background): “…detectors to determine the occurrence of individual objects for action recognition [25,36,30,38], some works integrate objects with the scene as context, e.g., Zheng et al. [43] learned context-based classifiers, which uses image content from both the foreground and the background as the context.”
Human Activity Recognition with Metric Learning (2008, background): “[4,5,6]; see also discussion in [7]), which suggests building appearance features for body configuration.”
Ordinal Depth Supervision for 3D Human Pose Estimation (2018, background): “This 2D-to-3D reconstruction ambiguity is mainly attributed to the binary ordinal depth relations of the joints (closer-farther) [45].”
Keep It SMPL: Automatic Estimation of 3D Human Pose and Shape from a Single Image (2016, methodology): “Lee and Chen [26] assume known limb lengths of a stick ﬁgure while Taylor [48] assumes the ratios of limb lengths are known.”
Monocular 3D Human Pose Estimation in the Wild Using Improved CNN Supervision (2016, background): “3D pose can be inferred from 2D pose through geometric and statistical priors [37, 59].”
Visual learning by imitation with motor representations (2005, methodology): “To reconstruct the 3D posture of the observed arm, we will follow the approach in [29], based on the orthographic camera and articulated arm models presented in Section 2.”
Fast pose estimation with parameter-sensitive hashing (2003, background): “In [18] 3D pose is recovered from the 2D projections of a number of known feature points on an articulated body.”
Sparse Representation for 3D Shape Estimation: A Convex Relaxation Approach (2015, neutral): “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.”
Detecting People Using Mutually Consistent Poselet Activations (2010, methodology): “Finding such parts requires extra annotation, and [14] introduced a new dataset, H3D, consisting of images of people annotated with 3D keypoints making use of Taylor’s algorithm [15].”
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”
Survey on Emotional Body Gesture Recognition (2018, methodology): “Model-based methods fit an expected model to the captured data, as an inverse kinematic problem [68], [69].”
Kinematic jump processes for monocular 3D human tracking (2003, methodology): “Skeletal reconstruction methods recover an interpretation tree of possible 3D joint positions, based on user-specified image joint positions [8, 17].”
Recognizing action at a distance (2003, methodology): “Alternatively we could go to the 3D skeleton from the 2D skeleton, as in [18].”
Sim2real transfer learning for 3D human pose estimation: motion to the rescue (2019, methodology): “One straightforward approach is to ‘lift’ 2D poses into 3D, using either dictionaries or direct regression [3, 44, 47, 64, 79, 85, 88, 93].”
Monocular 3D Human Pose Estimation Using Transfer Learning and Improved CNN Supervision (2016, methodology): “Here, model-based optimization is often required in addition to (iteratively) optimize the projection of a 3D human model to explain the 2D predictions.”