End-to-end depth from motion with stabilized monocular videos.
Clement PinardLaure ChevalleyAntoine ManzaneraDavid FilliatPublished in: CoRR (2018)
Keyphrases
- end to end
- depth from motion
- visual sensor
- optical flow
- image sequences
- video sequences
- visual odometry
- moving camera
- admission control
- congestion control
- video content
- motion estimation
- video data
- multipath
- video frames
- real time
- d scene
- transport layer
- computational complexity
- computer vision
- structure from motion
- moving objects