Latent-MVCNN: 3D Shape Recognition Using Multiple Views from Pre-defined or Random Viewpoints.
Qian YuChengzhuan YangHonghui FanHui WeiPublished in: Neural Process. Lett. (2020)
Keyphrases
- multiple views
- d objects
- multi view
- viewpoint
- single view
- multiple viewpoints
- multiple cameras
- point correspondences
- range images
- multiple images
- planar surfaces
- fundamental matrix
- object recognition
- three dimensional
- single image
- camera images
- camera calibration
- feature correspondences
- consistency constraints
- shape descriptors
- dynamic scenes
- photometric stereo
- pose estimation
- ground plane
- learning algorithm
- range data
- semi supervised
- overlapping views