TM2B: Transformer-Based Motion-to-Box Network for 3D Single Object Tracking on Point Clouds.
Anqi XuJiahao NieZhiwei HeXudong LvPublished in: IEEE Robotics Autom. Lett. (2024)
Keyphrases
- object tracking
- point cloud
- image sequences
- structure from motion
- moving objects
- particle filter
- video sequences
- robust tracking
- appearance model
- mean shift
- video surveillance
- kalman filter
- motion model
- visual tracking
- optical flow
- video objects
- camera motion
- motion estimation
- multi camera
- computer vision
- multiple cameras
- data association
- point sets
- sparse representation
- feature points
- moving camera
- multibody
- perspective images
- dynamic scenes
- graph cuts
- object recognition
- three dimensional