PerspectiveNet: 3D Object Detection from a Single RGB Image via Perspective Points.
Siyuan HuangYixin ChenTao YuanSiyuan QiYixin ZhuSong-Chun ZhuPublished in: NeurIPS (2019)
Keyphrases
- d objects
- pose parameters
- surface points
- physically plausible
- image contours
- feature points
- viewpoint
- image data
- geometric invariants
- object contours
- input image
- object recognition
- multi view
- object features
- image features
- three dimensional objects
- pose estimation
- object representation
- multi views
- multiple views
- surface properties
- range data
- planar objects
- single image
- surface patches
- image retrieval
- d mesh
- viewing direction
- cad models
- minimum bounding
- recognition and pose estimation
- object models
- static images
- complex background
- image segmentation
- image synthesis
- perceptual grouping
- vanishing points
- object classes
- cad model
- color information
- lighting conditions
- shape model
- point distribution
- hough transform
- keypoints
- three dimensional