Lightweight Monocular Depth Estimation via Token-Sharing Transformer.
Dong-Jae LeeJae Young LeeHyounguk ShonEojindl YiYeong-Hun ParkSung-Sik ChoJunmo KimPublished in: CoRR (2023)
Keyphrases
- lightweight
- depth estimation
- monocular images
- depth cues
- stereo vision
- depth map
- depth perception
- stereo matching
- depth information
- image sequences
- real scenes
- depth estimates
- scene understanding
- dynamic scenes
- feature matching
- super resolution
- stereo pair
- pose estimation
- disparity map
- stereo images
- model based pose estimation
- three dimensional
- belief propagation
- multi view
- vision system
- wireless sensor networks
- high resolution
- object recognition
- real time
- motion segmentation
- d scene
- object tracking
- active contours
- markov random field
- input image
- optical flow
- spatio temporal
- video sequences