Paper Info
Title
LinSEM: Linearizing segmentation evaluation metrics for medical images.
Abstract
•Mitigates the problem of evaluation metric having different meanings among objects.•Simulates actual segmentations covering qualities from excellent to unacceptable.•Constructs relationship between metric and acceptability score assigned by expert.•Linearized metrics show closer acceptability-meaning among objects.•Linearity of the metric is improved by LinSEM in an object-specific manner.
Year
DOI
Venue
2020
10.1016/j.media.2019.101601
Medical Image Analysis
Keywords
Field
DocType
Medical image segmentation,Evaluation metrics,Acceptability score,Linear relationship
Computer vision,Market segmentation,Pattern recognition,Segmentation,Sørensen–Dice coefficient,Linearity,Image segmentation,Jaccard index,Artificial intelligence,Hausdorff distance,Mathematics,Linearization
Journal
Volume
ISSN
Citations 
60
1361-8415
0
PageRank 
References 
Authors
0.34
0
5
Name
Order
Citations
PageRank
Jieyu Li134.80
Jayaram K. Udupa22481322.29
Yubing Tong39322.73
Lisheng Wang4209.08
D. A. Torigian58121.68