Informatics Report Series


Report   

EDI-INF-RR-1386


Related Pages

Report (by Number) Index
Report (by Date) Index
Author Index
Institute Index

Home
Title:Estimating the ground truth from multiple individual segmentations with application to skin lesion segmentation
Authors: Xiang Li ; Ben Aldridge ; Jonathan Rees ; Robert Fisher
Date:Jul 2010
Publication Title:Proc. Medical Image Understanding and Analysis
Publisher:BMVA Press
Publication Type:Conference Paper Publication Status:Published
Volume No:1 Page Nos:101-106
ISBN/ISSN:978-0-9566150-0-8
Abstract:
Having ground truth is critical for evaluating segmentation algorithms and finding the ground truth remains a hard problem. In this paper, three methods to estimate the ground truth for skin lesion segmentation using multiple manual results collected from different experts are proposed and compared. We also analyze the manual segmentations and discuss how to use them more effectively. We conclude that a voting policy produces a slightly better ground truth than the other two optimization based approaches. We pro- pose that a better ground truth should take into account different styles of segmentations.
Copyright:
2010 by The University of Edinburgh. All Rights Reserved
Links To Paper
No links available
Bibtex format
@InProceedings{EDI-INF-RR-1386,
author = { Xiang Li and Ben Aldridge and Jonathan Rees and Robert Fisher },
title = {Estimating the ground truth from multiple individual segmentations with application to skin lesion segmentation},
book title = {Proc. Medical Image Understanding and Analysis},
publisher = {BMVA Press},
year = 2010,
month = {Jul},
volume = {1},
pages = {101-106},
}


Home : Publications : Report 

Please mail <reports@inf.ed.ac.uk> with any changes or corrections.
Unless explicitly stated otherwise, all material is copyright The University of Edinburgh