Transformer-based out-of-distribution detection for clinically safe segmentation.
Mark S. GrahamPetru-Daniel TudosiuPaul WrightWalter Hugo Lopez PinayaJean-Marie U.-King-ImYee H. MahJames T. TeoHans Rolf JägerDavid WerringParashkev NachevSébastien OurselinM. Jorge CardosoPublished in: MIDL (2022)
Keyphrases
- level set
- reliable detection
- detection method
- image segmentation
- segmentation algorithm
- false positives
- landmark detection
- detection algorithm
- object detection
- boundary detection
- automatic detection
- fully automatic
- fuzzy logic
- cell segmentation
- image analysis
- energy function
- low signal to noise ratio
- segmentation method
- detection rate
- anomaly detection
- segmentation accuracy
- circular hough transform
- probability distribution
- image sequences
- pixel level
- bounding box
- false alarms
- shape prior
- automatically segmented
- neural network
- accurate localization
- foreground background separation
- automated detection
- fully unsupervised
- intensity variations
- automated analysis
- complex background
- video segmentation
- region growing
- active contours
- medical images
- multiscale