Query-based Temporal Fusion with Explicit Motion for 3D Object Detection.
Jinghua HouZhe LiuDingkang LiangZhikang ZouXiaoqing YeXiang BaiPublished in: NeurIPS (2023)
Keyphrases
- d objects
- physically plausible
- space time
- multi view
- object recognition
- partial matching
- viewpoint
- multiple views
- range data
- cad models
- pose estimation
- spatio temporal
- viewing direction
- motion model
- object classes
- object detection
- shape descriptors
- d mesh
- human motion
- image sequences
- motion estimation
- motion field
- image contours
- line drawings
- three dimensional
- three dimensional objects
- geometric models
- optical flow
- object features
- machine learning
- cad model
- video sequences
- object views
- pose normalization