OPA-3D: Occlusion-Aware Pixel-Wise Aggregation for Monocular 3D Object Detection.
Yongzhi SuYan DiGuangyao ZhaiFabian ManhardtJason R. RambachBenjamin BusamDidier StrickerFederico TombariPublished in: IEEE Robotics Autom. Lett. (2023)
Keyphrases
- d objects
- pixel wise
- pose estimation
- viewpoint
- multi view
- object recognition
- object detection
- three dimensional
- multiple views
- spatial information
- background subtraction
- partial occlusion
- appearance model
- image sequences
- pixel level
- template matching
- action recognition
- machine learning
- feature points
- image processing
- object classes
- foreground objects
- feature vectors
- computer vision