MonoJSG: Joint Semantic and Geometric Cost Volume for Monocular 3D Object Detection.
Qing LianPeiliang LiXiaozhi ChenPublished in: CVPR (2022)
Keyphrases
- d objects
- pose estimation
- object recognition
- geometric models
- viewpoint
- multi view
- partial matching
- topological information
- range data
- multiple views
- cad models
- three dimensional
- object parts
- object detection
- image sequences
- shape descriptors
- cad model
- image contours
- recognition and pose estimation
- d mesh
- object features
- surface points
- spin images
- object classes
- projective geometry
- stereo camera
- computer graphics
- pose normalization
- minimum bounding
- physically plausible
- optical flow