Measuring and evaluating ground truth for boundary detection in medical images

Irene Cheng; Carlos Flores-Mir; Paul Major; Anup Basu

doi:10.1109/IEMBS.2008.4650554

Measuring and evaluating ground truth for boundary detection in medical images

Annu Int Conf IEEE Eng Med Biol Soc. 2008:2008:5889-92. doi: 10.1109/IEMBS.2008.4650554.

Authors

Irene Cheng¹, Carlos Flores-Mir, Paul Major, Anup Basu

Affiliation

¹ NSERC Research Fellow at the Computer and Information Sciences Department, University of Pennsylvania, Canada. [email protected]

PMID: 19164057
DOI: 10.1109/IEMBS.2008.4650554

Abstract

In this paper we describe a methodology for measuring the ground truth for boundary detection methods in medical images. An interface is used to allow medical experts to outline boundaries of interest on images. Following this, the accuracy of the experts in repeating the ground truth detection process is evaluated. Finally, an approach is described for comparing the ground truth as provided by experts with actual boundaries detected automatically. Experimental results are presented to validate the proposed algorithms.

Publication types

Evaluation Study
Research Support, Non-U.S. Gov't

MeSH terms

Algorithms*
Artificial Intelligence*
Face / anatomy & histology*
Image Enhancement / methods
Image Interpretation, Computer-Assisted / methods*
Magnetic Resonance Imaging / methods*
Pattern Recognition, Automated / methods*
Reproducibility of Results
Sensitivity and Specificity