TY - GEN
T1 - Evaluation of image quality metrics for comparison of synchronization algorithms for cardiac cine MRI
AU - Nijm, G. M.
AU - Swiryn, S.
AU - Larson, A. C.
AU - Sahakian, A. V.
PY - 2008/12/1
Y1 - 2008/12/1
N2 - Cardiac MR image data is typically acquired over multiple cardiac cycles, thus necessitating the use of a synchronization algorithm. Comparison of such algorithms is challenging, as the MR environment distorts the signals that might be used to derive the gating signal for synchronization. In this study, we evaluated three cardiac synchronization algorithms by means of two full-reference image quality metrics and two no-reference metrics; we compared the resulting values of these metrics with subjective analysis by expert reviewers. Both full-reference metrics indicated that there was a statistically significant difference between the cardiac synchronization algorithms. However, both no-reference metrics indicated that there was no statistically significant difference between any of the synchronization algorithms, which corresponded with the subjective analysis. In conclusion, no-reference image metrics correlated well with expert analysis of cardiac MR images, while the full-reference metrics were most likely oversimplified which provided misleading results.
AB - Cardiac MR image data is typically acquired over multiple cardiac cycles, thus necessitating the use of a synchronization algorithm. Comparison of such algorithms is challenging, as the MR environment distorts the signals that might be used to derive the gating signal for synchronization. In this study, we evaluated three cardiac synchronization algorithms by means of two full-reference image quality metrics and two no-reference metrics; we compared the resulting values of these metrics with subjective analysis by expert reviewers. Both full-reference metrics indicated that there was a statistically significant difference between the cardiac synchronization algorithms. However, both no-reference metrics indicated that there was no statistically significant difference between any of the synchronization algorithms, which corresponded with the subjective analysis. In conclusion, no-reference image metrics correlated well with expert analysis of cardiac MR images, while the full-reference metrics were most likely oversimplified which provided misleading results.
KW - Cardiac magnetic resonance imaging
KW - Image quality metric
UR - http://www.scopus.com/inward/record.url?scp=70049088889&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=70049088889&partnerID=8YFLogxK
U2 - 10.1109/ICIP.2008.4712241
DO - 10.1109/ICIP.2008.4712241
M3 - Conference contribution
AN - SCOPUS:70049088889
SN - 1424417643
SN - 9781424417643
T3 - Proceedings - International Conference on Image Processing, ICIP
SP - 2260
EP - 2263
BT - 2008 IEEE International Conference on Image Processing, ICIP 2008 Proceedings
T2 - 2008 IEEE International Conference on Image Processing, ICIP 2008
Y2 - 12 October 2008 through 15 October 2008
ER -