Andrea Cellai

A high performance board for acquisition of 64-channel ultrasound RF data

The experimental test of new advanced ultrasound techniques frequently requires the acquisition of raw radiofrequency (RF) data from the individual elements of an array probe. The throughput rate and the total amount of these “pre-beamforming” data can be so high that only the most advanced research platforms are equipped to perform such acquisition, but may have difficulties to extend it over suitably long time intervals. In this paper, we report on the development of a high capacity (up to 36 GB) memory board which can be installed inside the Ultrasound Advanced Open Platform (ULAOP). 64 channel raw RF data can be acquired without interfering with the programmed real-time operation modalities of the system. Examples of application are discussed.

Embedded System for Real-Time Digital Processing of Medical Ultrasound Doppler Signals

Ultrasound (US) Doppler systems are routinely used for the diagnosis of cardiovascular diseases. Depending on the application, either single tone bursts or more complex waveforms are periodically transmitted throughout a piezoelectric transducer towards the region of interest. Extraction of Doppler information from echoes backscattered from moving blood cells typically involves coherent demodulation and matched filtering of the received signal, followed by a suitable processing module. In this paper, we present an embedded Doppler US system which has been designed as open research platform, programmable according to a variety of strategies in both transmission and reception. By suitably sharing the processing tasks between a state-of-the-art FGPA and a DSP, the system can be used in several medical US applications. As reference examples, the detection of microemboli in cerebral circulation and the measurement of wall _distension_ in carotid arteries are finally presented.

A new method for 2D-vector blood flow imaging based on unconventional beamforming techniques

Conventional Ultrasound (US) Doppler methods for blood flow imaging are limited to velocity estimations only in the axial direction, i.e. along the beam direction. Transverse oscillations (TO) methods extend blood investigations towards multidimensional estimates, and detailed descriptions of complex and fast blood flows are achievable by high frame-rate (HFR) imaging methods. In this work, TO are coupled with plane-waves (PWs) to reconstruct radio-frequency (RF) images with bi-directional oscillations in the pulse-echo field. The achieved RF images are exploited by a 2D phase-based displacement estimator to produce 2D-vector flow maps. A preliminary simulation study confirmed the capability of the method to produce the designed oscillations in the RF pulse-echo fields as well as the possibility to obtain 2D-vector maps with errors lower than 10% in many different conditions.

Frequency-domain high frame-rate 2D vector flow imaging

Conventional ultrasound flow imaging systems are limited to estimate only the axial component of blood velocity. In this work, a method to produce 2D vector Doppler maps is proposed and experimentally tested. The local displacements between consecutive high frame-rate (HFR) radio-frequency (RF) images are estimated in the frequency domain. Each image is subdivided in partially overlapped matching blocks, and the average local displacements in a block are calculated from the difference of spectral phases in consecutive frames. The method has been tested by simulations and experiments by using the ULA-OP research scanner. Preliminary in-vivo tests have been conducted and an example of the femoral vessels of a healthy volunteer is presented. The performance of the method are evaluated through the relative error bias and standard deviation, presenting values lower than 10% in standard conditions.

Multi-channel Raw-Data Acquisition for Ultrasound Research

Ultrasound is widely used in diagnostic applications where novel methods and techniques are continuously developed and proposed. The test of new techniques often requires the access to the raw echo-data saved from each of the multiple elements which compose the modern array probes. Given the high number of receiving elements and the high Analog-to-Digital sampling rate, tens of GB of data are typically generated in few seconds. Only a small number of research instruments are equipped to save raw data, but the saved quantity is often not sufficient. In this paper we describe a memory board that, coupled to the Ultrasound Advanced Open Platform (ULA-OP), can save up to 36 GB of data, sampled at 50 MHz, from 64 probe elements. Two novel applications developed by using the data from this board are discussed as well.

High frequency GaN-based pulse generator with active T/R switch circuit

High frequency ultrasound systems capable of high spatial resolution are commonly used for small biological structures imaging and non-destructive testing. For transducers operating around 50MHz central frequency, excitation pulses with amplitude of several tens of Volts and 100V/ns slew rate are needed. Different solutions have been proposed to isolate the receive Low Noise Amplifier (LNA) from the transmission signal (T/R switch). This paper proposes a high frequency transmit (TX) circuit with a T/R switch capable of overcoming the typical limitations of previous solutions.

ULA-OP: A Fully Open Ultrasound Imaging/Doppler System

Experimental research in Imaging/Doppler ultrasound frequently involves new transmission strategies, non-conventional beamforming techniques, custom data processing. This flexibility is not available on commercial equipment designed for clinical use but only on few existing research platforms which, on the other hand, are typically characterized by cumbersome and expensive electronics. In this paper, we present a novel ULtrasound Advanced Open Platform (ULA-OP), designed in the effort to overcome the aforementioned drawbacks. ULA-OP hardware and software are described in detail and the system capabilities are shown with two preliminary, non-conventional applications.