Hybridnet for depth estimation and semantic segmentation.
Dalila Sánchez-EscobedoXiao LinJosep R. CasasMontse PardàsPublished in: CoRR (2024)
Keyphrases
- depth estimation
- street scenes
- semantic segmentation
- stereo vision
- depth map
- stereo matching
- depth information
- scene understanding
- dynamic scenes
- stereo pair
- super resolution
- real scenes
- disparity map
- feature matching
- conditional random fields
- superpixels
- object categories
- vision system
- stereo images
- viewpoint
- specular reflection
- computer vision
- three dimensional
- scene classification
- weakly supervised
- image segmentation
- image sequences
- object classes
- data sets
- d scene
- space time
- higher order
- object detection
- ground truth
- object recognition