Real-Time Human Pose Tracking from Range Data

By Varun Ganapathi, Christian Plagemann, Daphne Koller, Sebastian Thrun. 2012

166 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

Space-Time Representation of People Based on 3D Skeletal Data: A Review (2016, neutral): “[61] ICP with physical constraints Depth image sequence 3D skeleton, real-time, 125fps, robust to self collision Plagemann et al.”
DART: Dense Articulated Real-Time Tracking (2014, background): “Another approach similar to ours is that of Ganapathi et al. [12], which uses range data to track articulated motion of human bodies by augmenting traditional ICP with free space information.”
Pose estimation for a partially observable human body from RGB-D cameras (2015, methodology): “Other model-based approaches [19], [20] produce impressive, accurate tracking and achieve”
Interactive Markerless Articulated Hand Motion Tracking Using RGB and Depth Data (2013, background): “• A near-realtime framework that captures hand motions with a level of precision and speed necessary for interactive applications.”
Metric Regression Forests for Correspondence Estimation (2015, methodology): “The pose is typically inferred using local optimization (Bregler et al. 2004; Brubaker et al. 2010; Stoll et al. 2011; Pons-Moll et al. 2011; Ganapathi et al. 2012) or stochastic search (Deutscher and Reid 2005; Gall et al.”
Articulated and Generalized Gaussian Kernel Correlation for Human Pose Estimation (2016, methodology): “Most generative methods involve explicit correspondence estimation in an ICP-like framework, where the pose and correspondence are iteratively and alternately updated, e.g. [7], [33].”
V2V-PoseNet: Voxel-to-Voxel Prediction Network for Accurate 3D Hand and Human Pose Estimation from a Single Depth Map (2017, methodology): “The ICP algorithm is commonly used for 3D body tracking [8, 13, 18, 22].”
A combined pose, object, and feature model for action understanding (2012, methodology): “We use the implementation of [8], which runs in real-time on an average laptop computer, as the basis of our system.”
Realtime and Robust Hand Tracking from Depth (2014, background): “Previous gradient based optimization for body tracking [27] and stochastic optimization for hand tracking [8, 9] cannot minimize the cost well, being either too sensitive to local optima or too slow in convergence.”
On Self-Contact and Human Pose (2021, background): “Similar ideas are adopted in [3, 10] where line segments are replaced with cylinders.”
VNect (2017, background): “Monocular Depth-based: The additional depth channel provided by RGB-D sensors has led to robust real-time pose estimation solutions [Baak et al. 2011; Ganapathi et al. 2012; Ma and Wu 2014; Shotton et al. 2013; Wei et al. 2012; Ye and Yang 2014] and the availability of low-cost devices has enabled a”
Towards Viewpoint Invariant 3D Human Pose Estimation (2016, methodology): “The EVAL dataset was recorded using the Microsoft Kinect camera at 30 fps. Similar to leave-one-out cross validation, we adopt a leave-one-out train-test procedure.”
Cascaded hand pose regression (2015, background): “Recent years have seen rapid progress and significant success of human body pose estimation [18, 29, 1, 19, 36] using consumer depth sensors.”
3D human action segmentation and recognition using pose kinetic energy (2014, background): “Going beyond RGB images and videos, the recent advent of inexpensive RGB-D sensors such as the Microsoft Kinect [1] that directly provide depth information has triggered an increasing interest in using depth images for pose estimation [30],[10] and action recognition [18],[31].”