Dense semantic labeling of sub-decimeter resolution images with convolutional neural networks.
Michele VolpiDevis TuiaPublished in: CoRR (2016)
Keyphrases
- convolutional neural networks
- image data
- image database
- image analysis
- three dimensional
- image resolution
- image matching
- input image
- image registration
- object recognition
- ground truth
- convolutional network
- edge detection
- semantic labels
- semantic categories
- image collections
- image classification
- manual labeling
- higher resolution
- web images
- highest resolution
- dense correspondence
- similarity measure
- image annotation
- image regions
- image retrieval
- test images
- segmentation method
- semantic information
- feature points
- image features
- active learning