Adaptable Deformable Convolutions for Semantic Segmentation of Fisheye Images in Autonomous Driving Systems.
Clément PlayoutOla AhmadFreddy LécuéFarida CherietPublished in: CoRR (2021)
Keyphrases
- semantic segmentation
- street scenes
- image database
- label transfer
- ground truth
- test images
- image set
- image retrieval
- input image
- autonomous driving
- three dimensional
- image features
- segmentation method
- pixel level
- computer vision algorithms
- multiple images
- object recognition
- image classification
- image understanding
- image regions
- image matching
- conditional random fields
- object categories
- superpixels
- keypoints
- single image
- graphical models
- probabilistic model
- image analysis