Real-Time Human Pose Recognition in Parts from Single Depth Images

  title={Real-Time Human Pose Recognition in Parts from Single Depth Images},
  author={Jamie Shotton and Andrew W. Fitzgibbon and Mat Cook and Toby Sharp and Mark Finocchio and Richard Moore and Alex Kipman and Andrew Blake},
  booktitle={Machine Learning for Computer Vision},
Moving Kinect-Based Gait Analysis with Increased Range
Results of the comparison of knee flexion, step length, and stride length with a software based method show that the moving Kinect system can accurately analyse these gait parameters.
Gait training assist system of a lower limb prosthetic visualizing muscle activation pattern using a color-depth sensor
A training assist system is developed, which estimates and visualizes an abductor muscle by using a color-depth sensor, and the effectiveness of this proposed algorithm has been confirmed.
Ensemble Learning for Visual Recognition
The author’s personal website is, and the book can be found here.
Estimating zero moment point and floor reaction force using visual markers
Validation of locomotive functions in elderly individuals is important in aging society. An algorithm to estimate the floor reaction forces on both feet were previously developed to evaluate the
Estimate Hand Poses Efficiently from Single Depth Images
A dedicated two-step regression forest pipeline is proposed: given an input hand depth image, step one involves mainly estimation of 3D location and in-plane rotation of the hand using a pixel-wise regression forest, which delivers final hand estimation by a similar regression forest model based on the entire hand image patch.
Frontal gait recognition from occluded scenes
Evaluating spatiotemporal interest point features for depth-based action recognition
One-Shot Person Re-identification with a Consumer Depth Camera
A comparison between two techniques for one-shot person re-identification from soft biometric cues based upon a descriptor composed of features provided by a skeleton estimation algorithm and a novel technique to warp the subject’s point cloud to a standard pose.
Depth sensor assisted real-time gesture recognition for interactive presentation
A Theory of Minimal 3D Point to 3D Plane Registration and Its Generalization
This paper provides a family of minimal solutions for the 3D-to-3D registration problem in which the3D data are represented as points and planes and identifies the degenerate configurations in such registrations.