SEMEDA: Enhancing segmentation precision with semantic edge aware loss.
Yifu ChenArnaud DapognyMatthieu CordPublished in: Pattern Recognit. (2020)
Keyphrases
- edge detection
- multiple scales
- segmentation algorithm
- image segmentation
- region growing
- edge information
- medical images
- segmentation method
- intermediate level vision
- shape prior
- segmentation accuracy
- level set
- multiscale
- semantic web
- segmented images
- image analysis
- object segmentation
- segmentation errors
- energy functional
- fully unsupervised
- semantic network
- fully automatic
- image regions
- semantic information
- energy function
- natural language
- semantic annotation
- object contours
- region segmentation
- piecewise constant
- high level