MV2MAE: Multi-View Video Masked Autoencoders.
Ketul ShahRobert CrandallJie XuPeng ZhouMarian GeorgeMayank BansalRama ChellappaPublished in: CoRR (2024)
Keyphrases
- multi view
- dynamic scenes
- free viewpoint
- single view
- depth video
- multiple views
- depth map
- stereo video
- virtual view synthesis
- camera calibration
- three dimensional
- d objects
- video sequences
- video content
- video data
- multiple cameras
- depth image based rendering
- video frames
- multi view stereo
- denoising
- semi supervised
- view synthesis
- depth information
- co training
- stereoscopic video
- free viewpoint video
- multi view images
- scene reconstruction
- multiple viewpoints
- motion vectors
- video surveillance
- range images
- input image
- surface reconstruction
- multi view reconstruction
- object recognition
- multi view video coding
- multi view clustering