In Defense of Pre-Trained ImageNet Architectures for Real-Time Semantic Segmentation of Road-Driving Images.
Marin OrsicIvan KresoPetra BevandicSinisa SegvicPublished in: CVPR (2019)
Keyphrases
- pre trained
- semantic segmentation
- image collections
- three dimensional
- input image
- image database
- image features
- image set
- test images
- image matching
- control signals
- image classification
- image annotation
- training examples
- small number
- bounding box
- probabilistic model
- conditional random fields
- feature points
- image content
- segmentation algorithm
- feature extraction
- object categories
- illumination conditions
- object recognition
- superpixels
- pixel level
- pascal voc
- image sequences
- image retrieval
- feature selection