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