Depth Estimation Matters Most: Improving Per-Object Depth Estimation for Monocular 3D Detection and Tracking.
Longlong JingRuichi YuHenrik KretzschmarKang LiCharles R. QiHang ZhaoAlper AyvaciXu ChenDillon CowerYingwei LiYurong YouHan DengCongcong LiDragomir AnguelovPublished in: CoRR (2022)
Keyphrases
- depth estimation
- monocular images
- stereo vision
- depth map
- stereo matching
- depth cues
- object tracking
- depth information
- scene understanding
- partial occlusion
- dynamic scenes
- stereo pair
- object detection
- real scenes
- super resolution
- disparity map
- feature matching
- real time
- depth estimates
- image sequences
- d scene
- visual tracking
- three dimensional
- appearance model
- stereo images
- surface reconstruction
- object boundaries
- human body
- particle filter
- complex scenes
- feature points
- vision system
- d objects
- high resolution
- video sequences
- event detection
- depth from defocus
- lighting conditions