TY - GEN
T1 - Constrained RF level interpolation for normalized cross correlation based speckle tracking
AU - Rebholz, Brandon
AU - Almekkawy, Mohamed
N1 - Publisher Copyright:
© 2020 IEEE.
PY - 2020/9/7
Y1 - 2020/9/7
N2 - Speckle tracking by normalized cross correlation (NCC) requires subsample accuracy for effective strain estimation. Without this, low displacement causes quantized images that result in unusable strain maps. Doppler methods work well on small displacement, however they operate very poorly on large displacement data where they are prone to aliasing. When manually palpating the medium, it is difficult to maintain a constant deformation rate and pressure, resulting in a mixture of large and small displacement data throughout a total data set. This paper presents the Constrained Radio Frequency method (CRF) of speckle tracking. This method uses the correlation gradients at the sample level to constrain the subsample search region such that the RF data can be interpolated to accurately estimate displacement with subsample precision. Simulated and experimental data sets are used to estimate displacement and strain to compare the CRF method to NCC speckle tracking and the Loupas algorithm doppler technique. The CRF is the most accurate method at estimating displacement and generates strain estimates with higher contrast to noise ratios (CNR) compared to the other methods tested.
AB - Speckle tracking by normalized cross correlation (NCC) requires subsample accuracy for effective strain estimation. Without this, low displacement causes quantized images that result in unusable strain maps. Doppler methods work well on small displacement, however they operate very poorly on large displacement data where they are prone to aliasing. When manually palpating the medium, it is difficult to maintain a constant deformation rate and pressure, resulting in a mixture of large and small displacement data throughout a total data set. This paper presents the Constrained Radio Frequency method (CRF) of speckle tracking. This method uses the correlation gradients at the sample level to constrain the subsample search region such that the RF data can be interpolated to accurately estimate displacement with subsample precision. Simulated and experimental data sets are used to estimate displacement and strain to compare the CRF method to NCC speckle tracking and the Loupas algorithm doppler technique. The CRF is the most accurate method at estimating displacement and generates strain estimates with higher contrast to noise ratios (CNR) compared to the other methods tested.
UR - http://www.scopus.com/inward/record.url?scp=85097882886&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85097882886&partnerID=8YFLogxK
U2 - 10.1109/IUS46767.2020.9251315
DO - 10.1109/IUS46767.2020.9251315
M3 - Conference contribution
AN - SCOPUS:85097882886
T3 - IEEE International Ultrasonics Symposium, IUS
BT - IUS 2020 - International Ultrasonics Symposium, Proceedings
PB - IEEE Computer Society
T2 - 2020 IEEE International Ultrasonics Symposium, IUS 2020
Y2 - 7 September 2020 through 11 September 2020
ER -