Self-supervised Multi-frame Monocular Depth Estimation with Pseudo-LiDAR Pose Enhancement.
Wenhua WuGuangming WangJiquan ZhongHesheng WangZhe LiuPublished in: ICRA (2023)
Keyphrases
- depth estimation
- multi frame
- monocular images
- super resolution
- high resolution
- depth map
- image sequences
- model based pose estimation
- low resolution
- stereo vision
- motion estimation
- point correspondences
- pose estimation
- image processing
- stereo matching
- image enhancement
- point cloud
- optical flow
- computationally efficient
- dynamic scenes
- high quality
- feature points
- motion analysis
- scene understanding
- depth information
- feature matching
- optic flow
- three dimensional
- stereo pair
- real scenes
- single image
- computer vision
- flow field
- viewpoint
- human body
- motion segmentation
- d scene
- video sequences
- specular reflection
- structure from motion
- human pose
- object tracking