TY - GEN
T1 - Strain imaging of arterial wall with translational motion compensation and error correction
AU - Hasegawa, Hideyuki
AU - Kanai, Hiroshi
PY - 2007
Y1 - 2007
N2 - To assess elastic properties, correlation techniques are widely used to measure the displacement and strain of the arterial wall caused by the heartbeat. However, the displacements estimated by the phase-sensitive correlation methods are biased when the frequency used for the displacement estimation is different from the center frequency of RF echo. One of reasons for the frequency variation is the interference of echoes. In the case of the arterial wall, the displacement due to radial translation is larger than that contributing to strain by a factor of 10 and, thus, the error resulting from the translational motion due to the mismatch between the frequency used for displacement estimation and the actual center frequency is not negligible compared with the minute displacement contributing to strain. In this study, a method is proposed in which the radial translation is removed prior to the calculation of complex correlation between echoes in two different frames to estimate the phase change between the echoes. The radial translation is removed by tracking the echo from the luminal interface of the wall because it is dominant compared with echoes from scatterers in the wall and is less affected by the interference. Using this procedure, the significant error resulting from the large translation can be greatly suppressed. After the removal of translation, an error correcting function based on complex correlation is introduced to further reduce the error due to the frequency mismatch. Accuracy improvement by the proposed method was validated using phantoms. As shown in the figure, the error in strain estimated by the proposed method was 12.0% from the theoretical strain profile, significantly smaller than that (23.7%) by the conventional method. Furthermore, in the in vitro experiments using extracted femoral arteries, the arterial wall containing calcified tissue showed very small strain in comparison with that almost homogeneously composed of fibrous tissue (mixture of smooth muscle and collagen). The means and the standard deviations of distensibility (calculated from strain and internal pressure) of fibrous and calcified tissues obtained for 7 sections of 5 femoral arteries were 2.42±2.1 and 0.35±0.50 MPa-1, respectively.
AB - To assess elastic properties, correlation techniques are widely used to measure the displacement and strain of the arterial wall caused by the heartbeat. However, the displacements estimated by the phase-sensitive correlation methods are biased when the frequency used for the displacement estimation is different from the center frequency of RF echo. One of reasons for the frequency variation is the interference of echoes. In the case of the arterial wall, the displacement due to radial translation is larger than that contributing to strain by a factor of 10 and, thus, the error resulting from the translational motion due to the mismatch between the frequency used for displacement estimation and the actual center frequency is not negligible compared with the minute displacement contributing to strain. In this study, a method is proposed in which the radial translation is removed prior to the calculation of complex correlation between echoes in two different frames to estimate the phase change between the echoes. The radial translation is removed by tracking the echo from the luminal interface of the wall because it is dominant compared with echoes from scatterers in the wall and is less affected by the interference. Using this procedure, the significant error resulting from the large translation can be greatly suppressed. After the removal of translation, an error correcting function based on complex correlation is introduced to further reduce the error due to the frequency mismatch. Accuracy improvement by the proposed method was validated using phantoms. As shown in the figure, the error in strain estimated by the proposed method was 12.0% from the theoretical strain profile, significantly smaller than that (23.7%) by the conventional method. Furthermore, in the in vitro experiments using extracted femoral arteries, the arterial wall containing calcified tissue showed very small strain in comparison with that almost homogeneously composed of fibrous tissue (mixture of smooth muscle and collagen). The means and the standard deviations of distensibility (calculated from strain and internal pressure) of fibrous and calcified tissues obtained for 7 sections of 5 femoral arteries were 2.42±2.1 and 0.35±0.50 MPa-1, respectively.
UR - http://www.scopus.com/inward/record.url?scp=48149084431&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=48149084431&partnerID=8YFLogxK
U2 - 10.1109/ULTSYM.2007.220
DO - 10.1109/ULTSYM.2007.220
M3 - Conference contribution
AN - SCOPUS:48149084431
SN - 1424413834
SN - 9781424413836
T3 - Proceedings - IEEE Ultrasonics Symposium
SP - 860
EP - 863
BT - 2007 IEEE Ultrasonics Symposium Proceedings, IUS
T2 - 2007 IEEE Ultrasonics Symposium, IUS
Y2 - 28 October 2007 through 31 October 2007
ER -