Joint Semantic Segmentation and Depth Estimation with Deep Convolutional Networks.
Arsalan MousavianHamed PirsiavashJana KoseckaPublished in: 3DV (2016)
Keyphrases
- street scenes
- depth estimation
- semantic segmentation
- stereo vision
- depth map
- stereo matching
- superpixels
- dynamic scenes
- scene understanding
- super resolution
- stereo pair
- conditional random fields
- depth information
- scene classification
- real scenes
- weakly supervised
- feature matching
- object classes
- multi view
- training data
- object categories
- active learning