Chat: Cascade Hole-Aware Transformers with Geometric Spatial Consistency for Accurate Monocular Endoscopic Depth Estimation.
Ming WuHao QiWenkang FanSunkui KeHui-Qing ZengYinran ChenXiongbiao LuoPublished in: ICASSP (2024)
Keyphrases
- depth estimation
- depth estimates
- monocular images
- depth map
- stereo vision
- depth cues
- depth information
- stereo matching
- depth perception
- dynamic scenes
- real scenes
- image sequences
- stereo pair
- scene understanding
- space time
- feature matching
- high quality
- three dimensional
- blur estimation
- pose estimation
- super resolution
- spatio temporal
- geometric constraints
- multi view
- d scene
- video sequences
- structure from motion
- multiple views
- human body
- ground truth
- viewpoint
- real time