FCN-Boosted Historical Map Segmentation with Little Training Data.
Josef BalounLadislav LencPavel KrálPublished in: ICDAR (1) (2023)
Keyphrases
- training data
- image segmentation
- segmentation algorithm
- region growing
- segmentation method
- multiscale
- level set
- shape prior
- training set
- medical images
- edge detection
- data sets
- supervised learning
- parametric representation
- object segmentation
- test set
- decision trees
- motion segmentation
- fully automatic
- neural network
- aerial photographs
- fully unsupervised
- training dataset
- learned from training data
- maximum a posteriori
- energy function
- training samples
- active contours
- classification accuracy
- image analysis
- computer vision
- learning algorithm