Multi-view Action Recognition via Directed Gromov-Wasserstein Discrepancy.
Hoang-Quan NguyenThanh-Dat TruongKhoa LuuPublished in: CoRR (2024)
Keyphrases
- multi view
- action recognition
- view invariant
- single view
- human actions
- multiple views
- camera calibration
- three dimensional
- computer vision
- multiple cameras
- d objects
- activity recognition
- depth map
- surface reconstruction
- body parts
- human pose
- semi supervised
- object recognition
- view synthesis
- human activities
- motion capture
- spatio temporal
- visual hull
- multiple viewpoints
- object detection