## Abstract

A method for estimating the velocity spectrum for a fully transverse flow at a beam-to-flow angle of 90is described. The approach is based on the transverse oscillation (TO) method, where an oscillation across the ultrasound beam is made during receive processing. A fourth-order estimator based on the correlation of the received signal is derived. A Fourier transform of the correlation signal yields the velocity spectrum. Performing the estimation for short data segments gives the velocity spectrum as a function of time as for ordinary spectrograms, and it also works for a beam-to-flow angle of 90. The approach is validated using Field II simulations. A 3 MHz convex array with lambda pitch is modeled. The transmit focus is at 200 mm and 2 times 32 elements are used in receive. A dual-peak Hamming apodization with a spacing of 96 elements between the peaks is used during receive beamforming for creating the lateral oscillation. Pulsatile flow in a femoral artery

placed 40 mm from the transducer is simulated for one cardiac cycle using the Womersly-Evan’s flow model. The bias of the mean estimated frequency is 13.6% compared to the true velocity and the mean relative std is 14.3%. This indicates that the new estimation scheme can reliably find the spectrum at 90, where a traditional estimator yields zero velocity. Measurements have been conducted with the SARUS experimental scanner and a BK 8820e convex array transducer (BK Medical, Herlev, Denmark). A CompuFlow 1000 (Shelley Automation, Inc, Toronto, Canada) generated the artificial femoral artery flow in the phantom. It is demonstrated that the transverse spectrum can be found from the measured data. The estimated spectra degrade when the angle is different from 90, but are usable down to 60-70. Below this angle the traditional spectrum is best and should be used. The conventional approach can automatically be corrected for angles from 0-70to give a fully quantitative velocity spectrum without operator intervention.

placed 40 mm from the transducer is simulated for one cardiac cycle using the Womersly-Evan’s flow model. The bias of the mean estimated frequency is 13.6% compared to the true velocity and the mean relative std is 14.3%. This indicates that the new estimation scheme can reliably find the spectrum at 90, where a traditional estimator yields zero velocity. Measurements have been conducted with the SARUS experimental scanner and a BK 8820e convex array transducer (BK Medical, Herlev, Denmark). A CompuFlow 1000 (Shelley Automation, Inc, Toronto, Canada) generated the artificial femoral artery flow in the phantom. It is demonstrated that the transverse spectrum can be found from the measured data. The estimated spectra degrade when the angle is different from 90, but are usable down to 60-70. Below this angle the traditional spectrum is best and should be used. The conventional approach can automatically be corrected for angles from 0-70to give a fully quantitative velocity spectrum without operator intervention.

Original language | English |
---|---|

Title of host publication | Proceedings of IEEE International Ultrasonics Symposium |

Publisher | IEEE |

Publication date | 2013 |

Pages | 631-634 |

ISBN (Print) | 9781467356862 |

DOIs | |

Publication status | Published - 2013 |

Event | 2013 IEEE International Ultrasonics Symposium - Prague Convention Center , Prague, Czech Republic Duration: 21 Jul 2013 → 25 Jul 2013 http://ewh.ieee.org/conf/uffc/2013/ |

### Conference

Conference | 2013 IEEE International Ultrasonics Symposium |
---|---|

Location | Prague Convention Center |

Country/Territory | Czech Republic |

City | Prague |

Period | 21/07/2013 → 25/07/2013 |

Internet address |

## Keywords

- Fields, Waves and Electromagnetics
- Arteries
- Blood
- Estimation
- Oscillators
- Spectrogram
- Ultrasonic imaging
- Ultrasonic variables measurement