3DPPE: 3D Point Positional Encoding for Transformer-based Multi-Camera 3D Object Detection.
Changyong ShuJiajun DengFisher YuYifan LiuPublished in: ICCV (2023)
Keyphrases
- d objects
- multi camera
- multiple cameras
- multi view
- pose estimation
- viewpoint
- multiple views
- control points
- object recognition
- three dimensional
- shape descriptors
- range data
- object tracking
- minimum bounding
- object classes
- viewing direction
- point distribution
- computer vision
- surveillance system
- object detection
- multiscale
- machine learning
- multi sensor
- video surveillance
- event detection
- data fusion
- ground plane
- multiresolution