Detaching and Boosting: Dual Engine for Scale-Invariant Self-Supervised Monocular Depth Estimation.
Peizhe JiangWei YangXiaoqing YeXiao TanMeng WuPublished in: IEEE Robotics Autom. Lett. (2022)
Keyphrases
- scale invariant
- depth estimation
- depth map
- stereo vision
- image sequences
- keypoints
- natural images
- stereo matching
- scale space
- affine invariant
- dynamic scenes
- depth information
- super resolution
- scene understanding
- real scenes
- pose estimation
- multiscale
- invariant features
- local binary pattern
- feature matching
- image structure
- texture images
- stereo pair
- feature selection
- disparity map
- low resolution
- input image
- image analysis
- high quality
- machine learning
- stereo images
- background subtraction
- d scene
- multi view
- optical flow
- training data
- three dimensional