Active-learning-driven deep interactive segmentation for cost-effective labeling of crop-weed image data.
Freddy SikouonmeuMartin AtzmuellerPublished in: GIL Jahrestagung (2023)
Keyphrases
- cost effective
- active learning
- interactive segmentation
- image data
- interactive image segmentation
- automatic segmentation
- user interaction
- graph cuts
- image segmentation
- energy minimization
- low cost
- fully automatic
- active learning strategies
- machine learning
- random sampling
- image quality
- learning algorithm
- cost effectiveness
- loopy belief propagation
- supervised learning
- segmentation result
- computer vision
- training examples
- accurate segmentation
- semi supervised
- labeled data
- shape prior
- object segmentation
- multiscale
- image processing
- training data
- image content
- training set
- mr images
- energy function
- user interface
- em algorithm
- graphical models