Efficient ConvNet-based marker-less motion capture in general scenes with a low number of cameras.
Ahmed ElhayekEdilson de AguiarArjun JainJonathan TompsonLeonid PishchulinMykhaylo AndrilukaChristoph BreglerBernt SchieleChristian TheobaltPublished in: CVPR (2015)
Keyphrases
- motion capture
- markerless
- human motion capture
- human body
- human motion
- motion tracking
- multiple cameras
- human movement
- human computer interaction
- markerless motion capture
- multi camera
- pose estimation
- computer vision
- augmented reality
- multi view
- motion capture data
- camera calibration
- computer graphics
- viewpoint
- feature space