Are we using appropriate segmentation metrics? Identifying correlates of human expert perception for CNN training beyond rolling the DICE coefficient.
Florian KoflerIvan EzhovFabian IsenseeFabian BalsigerChristoph BergerMaximilian KoernerJohannes C. PaetzoldHongwei LiSuprosanna ShitRichard McKinleySpyridon BakasClaus ZimmerDonna P. AnkerstJan KirschkeBenedikt WiestlerBjoern H. MenzePublished in: CoRR (2021)
Keyphrases
- human experts
- image segmentation
- post processing
- segmentation algorithm
- domain experts
- level set
- segmentation method
- training set
- domain specific knowledge
- medical images
- training examples
- region growing
- cellular neural networks
- data sets
- training samples
- classification accuracy
- image segmentation algorithms
- perceptual information