Joint Semantic Segmentation and Depth Estimation with Deep Convolutional Networks.
Arsalan MousavianHamed PirsiavashJana KoseckaPublished in: CoRR (2016)
Keyphrases
- street scenes
- depth estimation
- semantic segmentation
- stereo vision
- depth map
- stereo matching
- depth information
- dynamic scenes
- conditional random fields
- scene understanding
- superpixels
- super resolution
- multi view
- scene classification
- stereo pair
- feature matching
- three dimensional
- real scenes
- disparity map
- probabilistic model
- computer vision
- stereo images
- specular reflection
- dynamic programming
- vision system
- weakly supervised
- image sequences
- segmentation algorithm