MoDAR: Using Motion Forecasting for 3D Object Detection in Point Cloud Sequences.
Yingwei LiCharles R. QiYin ZhouChenxi LiuDragomir AnguelovPublished in: CoRR (2023)
Keyphrases
- d objects
- point cloud
- real world objects
- cad model
- viewpoint
- object recognition
- physically plausible
- structure from motion
- multi view
- range data
- three dimensional
- point sets
- object detection
- pose estimation
- image sequences
- laser scanner
- d mesh
- surface reconstruction
- shape descriptors
- multiple views
- space time
- camera motion
- visual hull
- object classes
- stereo camera
- object surface
- moving objects
- feature vectors
- shape matching
- video sequences
- depth map
- missing data
- computer vision