Academic
Publications
Real time motion capture using a single time-of-flight camera

Real time motion capture using a single time-of-flight camera,10.1109/CVPR.2010.5540141,Varun Ganapathi,Christian Plagemann,Daphne Koller,Sebastian Th

Real time motion capture using a single time-of-flight camera   (Citations: 12)
BibTex | RIS | RefWorks Download
Markerless tracking of human pose is a hard yet relevant problem. In this paper, we derive an efficient filtering algorithm for tracking human pose using a stream of monocular depth images. The key idea is to combine an accurate generative model - which is achievable in this setting using programmable graphics hardware - with a discriminative model that provides data-driven evidence about body part locations. In each filter iteration, we apply a form of local model-based search that exploits the nature of the kinematic chain. As fast movements and occlusion can disrupt the local search, we utilize a set of discriminatively trained patch classifiers to detect body parts. We describe a novel algorithm for propagating this noisy evidence about body part locations up the kinematic chain using the unscented transform. The resulting distribution of body configurations allows us to reinitialize the model-based search. We provide extensive experimental results on 28 real-world sequences using automatic ground-truth annotations from a commercial motion capture system.
Conference: Computer Vision and Pattern Recognition - CVPR , pp. 755-762, 2010
Cumulative Annual
View Publication
The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.
    • ...The task has recently been greatly simplified by the introduction of realtime depth cameras [16, 19, 44, 37, 28, 13]...
    • ...Our focus on per-frame initialization and recovery is designed to complement any appropriate tracking algorithm [7, 39, 16, 42, 13] that might further incorporate temporal and kinematic coherence...
    • ...We show both qualitative and quantitative results on several challenging datasets, and compare with both nearest-neighbor approaches and the state of the art [13]...
    • ...We also evaluate on the real depth data from [13]...
    • ...While hierarchical matching [14] is faster, one would still need a massive exemplar set to achieve comparable accuracy. Comparison with [13]...
    • ...The authors of [13] provided their test data and results for direct comparison...
    • ...(a) Comparison with nearest neighbor matching. (b) Comparison with [13]...
    • ...Even without the kinematic and temporal constraints exploited by [13], our algorithm is able to more accurately localize body joints...

    Jamie Shottonet al. Real-Time Human Pose Recognition in Parts from a Single Depth Image

    • ...We evaluate on several datasets including the MSRC dataset of 5000 synthetic depth images [18] and the Stanford dataset of real depth images [8], obtaining state of the art results.,Beyond predicting joint positions more accurately, our algorithm also makes predictions faster, running at 200 fps compared to the 50 fps achieved by [18]. See also Fig. 7c. Ganapathi et al. [8].,This closely matches the results of [18] (0.947), and surpasses the result of [8] (0.898) which additionally exploited temporal and kinematic constraints...

    Ross Girshicket al. Efficient regression of general-activity human poses from depth images

    • ...Nonetheless, ToF data suffers from noise and estimating human full-body pose remains a difficult problem [7]...
    • ...Without modifying the interest point detection technique, the authors add in [7] a pose estimation method embedded in a Bayesian tracking framework...

    Loren Arthur Schwarzet al. Estimating human 3D pose from Time-of-Flight images based on geodesic ...

    • ...Many algorithms have been proposed to address the problem of pose estimation and motion capture from range images [5, 7, 11, 16]...
    • ...Ganapathi et al. [5] present a filtering algorithm to track human poses using a stream of depth images captured by a TOF camera...

    Xia Luet al. Human detection using depth information by Kinect

    • ...First review is from Zhu et al. [20]. Second is from Ganapathi et al. [21]...
    • ...Firstly, Ganapathi et al. [21] actually predicts the future about motion capture technology become convenient, cheap, and applicable in natural interaction environment...
    • ...In their presented system, estimation for joint angles of a 48 degree-of-freedom (DOF) human model by requiring four to ten frame per second (FPS) [21]...
    • ...Based on Ganapathi et al. [21] were highlighted two contributions inside their research paper...
    • ...Second contribution is a definition of a smooth likelihood function and a means of implementing it on readily available graphics hardware (GPUs) efficiently in order to obtain near real-time performance [21]...
    • ...Ganapathi et al. [21] list the goals as shown: (1) our proposed system is able to estimate the pose and configuration of a human over time using only a stream of depth images, (2) proposing candidates using EP on detected body parts significantly improves performance over just doing local hill-climbing, (3) the smoothed energy function outperforms the typically used pixel-wise energy function, and (4) the system runs close to real-time...
    • ...The algorithm detects moving regions by subtracting two consecutive frames from each other and highlighting the difference [21]...
    • ...2010 Zhu et al. [20] SwissRanger SR-3000 [23] Whole body Bayesian Framework 2010 Ganapathi et al. [21] SwissRanger SR-4000 [26] Whole body HC+EP...

    Mohd Kufaisal bin Mohd Sidiket al. A Study on Natural Interaction for Human Body Motion Using Depth Image...

Sort by: