Monocular Estimation of Translation, Pose and 3D Shape on Detected Objects using a Convolutional Autoencoder.
Ivar PerssonMartin AhrnbomMikael G. NilssonPublished in: VISIGRAPP (5: VISAPP) (2022)
Keyphrases
- d objects
- pose estimation
- pose parameters
- pose tracking
- multi view
- object geometry
- object pose
- monocular images
- human pose
- object shape
- viewpoint
- image sequences
- position and orientation
- pose recovery
- shape estimation
- human body
- moving objects
- object recognition
- object classes
- curved surfaces
- partial occlusion
- calibrated cameras
- computer vision
- object representation
- restricted boltzmann machine