GTA: A Geometry-Aware Attention Mechanism for Multi-View Transformers.
Takeru MiyatoBernhard JaegerMax WellingAndreas GeigerPublished in: ICLR (2024)
Keyphrases
- multi view
- attention mechanism
- three dimensional
- multi view stereo
- view dependent
- single view
- visual attention
- multiple views
- depth map
- d objects
- multiple cameras
- saliency map
- camera calibration
- range images
- semi supervised
- visual attention model
- surface reconstruction
- viewpoint
- multi view images
- multi views
- multi view face detection
- free viewpoint
- multiple viewpoints
- higher order
- scene geometry
- light field
- geometric constraints
- image processing