End-to-End Trainable Deep Active Contour Models for Automated Image Segmentation: Delineating Buildings in Aerial Imagery.
Ali HatamizadehDebleena SenguptaDemetri TerzopoulosPublished in: CoRR (2020)
Keyphrases
- end to end
- aerial imagery
- active contour model
- aerial images
- active contours
- image segmentation
- elevation data
- segmentation method
- level set method
- lidar data
- curve evolution
- energy minimization
- region growing
- gvf snake
- man made structures
- shape prior
- energy functional
- gradient vector flow
- level set
- graph cuts
- urban areas
- multiscale
- watershed segmentation
- markov random field
- energy function
- line segments
- building facades
- global minimizers
- segmentation result
- object segmentation
- point cloud
- image processing
- high resolution
- computer vision
- segmentation algorithm
- region merging
- pairwise
- image data
- high quality
- urban environments