3-D model-based tracking of humans in action: a multi-view approach

By Dariu M. Gavrila, Larry S. Davis. 1996

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

Histogram of Oriented Principal Components for Cross-View Action Recognition (2014, methodology): “some techniques infer 3D scene structure and use geometric transformations to achieve view invariance [15], [16], [17], [18], [19].”
HOPC: Histogram of Oriented Principal Components of 3D Pointclouds for Action Recognition (2014, neutral): “representations are proposed such as view invariant spatio-temporal features [2,20,22,23,25,33]. Some methods infer the 3D scene structure and use geometric transformations to achieve view invariance [5,9,15,26,39]. Another approach is to ﬁnd a view independent latent space [7,8,12,14] in which feat”
Discriminative virtual views for cross-view action recognition (2012, background): “…scene structure so that the derived action descriptors can be adapted from one view to another through geometric reasoning [29, 26, 15, 10, 4], while another is to search for spatio-temporal features of a video sequence that are insensitive to changes in view angle [17, 21, 23, 22, 3,…”
Inferring 3D body pose from silhouettes using activity manifold learning (2004, background): “Such constraints are explicitly exploited to recover the body configuration and motion in model-based approaches, such as [12, 10, 21, 20, 9, 14, 26], through explicitly specifying articulated models of the body parts, joint angles and their kinematics (or dynamics) as well as models for camera…”
Human motion analysis: a review (1997, methodology): “Since the authors in [19] claimed that their method was general to multiple calibrated cameras, Gavrila and Davis [36, 37 ] extended it to four calibrated cameras for 3D human body modeling.”
Recent developments in human motion analysis (2003, background): “Compared with 2-D approaches, 3-D approaches are more effective for accurate estimation in physical space, effective handling of occlusion, and the high-level judgment between various complex human movements such as wandering around, shaking hands and dancing [72,92-94].”
Distance-based outliers: algorithms and applications (2000, methodology): “There are a number of techniques used for tracking people [14, 16, 19].”
Human Body Model Acquisition and Tracking Using Voxel Data (2003, methodology): “Gavrila and Davis (1996) used a human body model composed of tapered super-quadrics to track (multiple) people in 3D.”
Tracking People by Learning Their Appearance (2007, unknown): “whicharetypicallyvariantsofdynamicprogramming[4],[5], [6], kalman filtering [7], [ 8 ], [9], or particle filtering [10], [11], [12],[13],[14],[15].Oneusestheposeinthecurrentframeand”
Multi-view 3D Human Pose Estimation in Complex Environment (2011, methodology): “…of the actual values used in the experiments, see Appendix A. 3.2 3D Human Shape Model Our 3D upper body model uses tapered super-quadrics as body part primitives (Gavrila and Davis 1996); this yields a good trade-off between desired accuracy and model complexity (i.e. number of parameters).”
Articulated body motion capture by annealed particle filtering (2000, background): “Gavrila [4] does just this when he uses what he terms search space decomposition.”
A Survey of Computer Vision-Based Human Motion Capture (2001, methodology): “1983 Hogg [57] Cylinders 14 Body Edges 2 1 2 1984 Hogg [58] Cylinders 14 Body Edges 2 1 2 1985 Lee [82] Stick-Figure 17 Body Joints 3 1989 Attwood [6] Cylinders 16 Body Joints 3 1991 Wang [139] Cylinders 2 Leg Motion 3 1991 Yamamoto [151] CAD Model 4 U Body Motion 2 1 2 1992 Lee [83] Stick-Figure 17”
Detection and location of people in video images using adaptive fusion of color and edge information (2000, unknown): “Many authors have developed methods of detecting people in images [1] [2] [3] [4] [6] [7].”
A survey on activity recognition and behavior understanding in video surveillance (2013, background): “Gavrilla [48]”
Real-Time Visual Tracking of Complex Structures (2002, methodology): “Gavrila and Davis [ 15 ] achieve this by matching edges in the image with those of an appearance model using distance transforms, Delamarre and Faugeras [10] force the appearance model to fit within silhouettes computed from multiple images, while Bregler and Malik [5] compute adaptive support maps ”
Detailed Human Shape and Pose from Images (2007, background): “[8]) and various deformable models such as superquadrics [9, 14, 20] and free-form surface patches [17].”
Model-based 3D tracking of an articulated hand (2001, methodology): “A direct search of the model parameters in order to match the model projection with the images was presented in [12].”
Neural Body Fitting: Unifying Deep Learning and Model Based Human Pose and Shape Estimation (2018, methodology): “Early models were based on geometric primitives [34, 11, 39, 50, 54], while more recent ones are estimated from 1000s of scans of real people, and are typically parameterized by separate body pose and shape components [5, 13, 28, 71, 41], with few exceptions, e.”